Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
This PDF is a selection from an out-of-print volume from the National Bureau of Economic Research Volume Title: Explorations in Economic Research, Volume 2, number 4 Volume Author/Editor: NBER Volume Publisher: NBER Volume URL: http://www.nber.org/books/meye75-2 Publication Date: October 1975 Chapter Title: The U.S. Railroad Industry in the Post-World War II Period: A Profile Chapter Author: John R. Meyer, Alexander L. Morton Chapter URL: http://www.nber.org/chapters/c9072 Chapter pages in book: (p. 449 - 501) 1 JOHN R. MEYER National Bureau of Economic Research and Harvard University ALEXANDER 1. MORTON National Bureau of Economic Research and Harvard University The U.S. Railroad Industry in the Post-World War II Period: A Profile 'I y V ABSTRACT: 0 0 e C a) Railroading has been a troubled industry for half a century. The troubles have approached the crisis stage once again during the past few years. Much of the railroad system in the industrial heartland of the nationeast of Chicago and north of the Potomac and Ohio riversis in bankruptcy. These bankruptcies, furthermore, are not yielding to the traditional solution of financial reorganizations. The rate of return on the investment of Class railroads even in the "prosperous" south and west has averaged only about 3.6 percent per year during the past six years. The recent surge in railroad traffic, causing ton-miles to attain all-time peaks, has not generated a corresponding improvement in profits, thus causing railroad managements and analysts to speak of their "profitless boom"! This gives rise to fears C d 0 '5 h NOTE: This study draws on research performed for the Task Force on Raitroad Productivity, of which the authors were members. The Task Force was co-sponsored by the National Commission on Productivity, the Council of Economic Advisers, and the Office of Science and Telhnotogy. Other members of the Task Force included Robert E. Gallamore, tawrence Gotdmuntz, George Hilton, lohn W. Ingram, A. Stheffer Lang, Man K. McAdams, Thomas G. Moore, John Morrissey, James R. Nelson, and Robert 0. lollison. The final report of the Task Force, Jmproving Radrnad Productivity, was released by the National Commission on Productivity in November 1973. The authors thank Atherton Bean, Solomon Fabricant, and Theodore 0. Yntema of the Directors' reading committee and Nathaniel Goldfinger and Mark Roberts for their valuable comments, and Miss Sally Aldrich 4) who typed this study with unselfish dedication 449 450 John R. Meyer and Alexander L Morton that another economic recession may send more railroads into bankruptcy. ¶ The origins of the railroad problem reside in the simple tact that the economy is expanding in areas that do not typically produce much new traffic of a type suited to conventional rail transport. Competition from trucking, other specialized modes of transport, and transport alternatives (such as high voltage transmission of electricity, synthetic materials, and recycling of waste materials) are further circumscribing the demand for rail freight service. ¶ The railroads have also suffered productivity problems, caused in part by these shifts in the freight market. Employment in the industry has declined by nearly two thirds during the postwar period, enabling the industry to record gains in labor productivity as great or greater than the average for private industry. However, to achieve this reduction in labor inputs, the industry has had to maintain or slightly increase its employment of capital; as a result, railroad capital productivity, the ratio of output to capital inputs, has shown no growth or has even declined slightly. Total or multiple factor productivity measures that combine labor and capital inputs suggest that railroads' overall productivity gains have been no higher and perhaps lower than the average for private industry. [11 INTRODUCTION: BASIC FINANCIAL TRENDS Railroading in the United States has been a troubled industry for half a century. In this paper an attempt is made to document what those troubles have been and why they may have occurred. To the degree possible, the emphasis throughout is on the quantitatje facts of the United States rail industry's situation during the post-World War II years. These facts are supplemented as necessary by qualitative considerations and basic historical trends within the industry and the surrounding economic environment. Since railroads in the United States are still private firms in a market economy, it seems sensible to start with the profitability or earnings record of the enterprises involved. Table 1 shows the earnings of Class I railroad companies from railroad operations and from other sources for the last four decades.1 A distinct long-run decline in net railway operating income has occurred and been partially offset by an upward trend in income from other sources. The gradual upward trend in "other income" reflects the accelerated development of the extensive real estate and mineral rights of railroads and the diversification of railroad The profitability of rail operations alone companies into other industries. (exclusive of other income) can be measured by what is called net railway operating income (NROI), shown in column 2 of Table 1 . In addition to the rather sharp downward course of net 1252 589 620 1929 1939 1946 1947 1948 1949 1128 1068 922 762 745 584 538 1955 1956 1957 1958 1959 1960 1961 1952 1953 1954 1951 1040 943 1078 1109 874 1950 1002 686 781 Other Net Railway Operating Incomea Year 346 322 276 289 333 315 271 266 273 289 309 277 210 229 235 254 161 360 Incomea (3) (2) 930 860 1062 1399 1344 1212 1095 1305 1216 1367 1419 1151 941 750 830 1010 1237 1512 Tota Incomea (4) 27,181 27,474 26,851 27,234 27,688 27,538 27,431 26,670 26,670 26263 24,592 25,519 2.13 1.97 4.22 3.95 3.36 2.76 2.72 4.26 3.76 4.16 4.19 3.28 2.88 4.31 23,624 24,028 22,891 2.56 2.75 3.44 53() 23,017 22,549 23868 (%) col. 1col. 2 (6) Crude Rate of Return on Net Investment (5) Net Investment in Transportation Property (billions) Income (After Taxes) and Rate of Return on Net Investment in Transportation Property of Class I Railroads, 1929-1972 (1) TABLE 1 (2) 482 422 395 365 399 458 500 505 326 330 369 Incomea Other (3) 968 1115 1230 1327 1445 1134 1177 1160 1 1051 136 87 lncome Total (4) federal Income taxes) but before fixed charges. All figures in current dolla#s 486 696 835 962 1046 676 678 655 726 806 818 Net Railway Operating Incomea (concluded) "AU fgurs are after taxes (inc!udng 1971 1972 1970 1965 1966 1967 1968 1969 1962 1963 1964 Year (1) TABLE 1 (5) 28,186 28,100 28,280 26,319 27,322 27,733 27,668 27,892 25,773 25,989 25858 Net Investment in Transport ion Property (billions) (6) 2.47 2.95 1.73 3.69 3.90 2.46 2.44 2.36 3.16 1.12 2.74 (%) col. 1-co), 2 Crude Rate of Return on Net Investment The U.S. Railroad Industry in the Post-World War II Period: A Profile 453 railway operating income over the postwar period, there is a strong in-phase cyclical component. Most significantly, operating income hit successively lower levels during the recessions of 1954, 1958, 1961, and 1970. Net investment in transportation property, including cash and materials inventories,3 is shown in column 5 of Table i and the rate of return on this invest- ment (column 2 ± column 5) is shown in column 6. Compared to more recent years, 1972 was a "good year" for the railroad industry. Yet the return of 2.95 percent puts railroads next to the last among 62 industries ranked according to rate of return on net worth in 1972; their return on equity of 3 percent is less than one-third that for all private industry (10.5 percent).' In only five of the years since 1947 has rail rate of return exceeded 4 percent, and the most recent of those years is 1955. Since 1955 the rate of return has exceeded 3 percent in only six out of 16 years, and the most recent of those years is 1966. Because net investment in transportation property displays a modest growth over the postwar period, the rate of return exhibits an even sharper downward trend than net railway operating income.6 The earnings performance of the railroad industry portrayed in the foregoing statistics is, of course, considerably influenced by the low profits or actual losses of the northeast railroads. For example, the losses of six deficit eastern carriers reduced net income of the rail industry by $386 million in 1970 and by $321 million in 1971. Table 2, showing the rate of return by district, reveals that it is primarily the mounting problems of ailing railroads in the northeast that have caused the rate of return for the entire industry to decline over the postwar period. However, a low rate of return on investment is not unique to eastern district roads. The railroads of the southern district have maintained a rate of return in the range of only 4-5 percent and 1972 was the first year since 1956 that a 5 percent return was achieved. Western district roads have maintained a return in the 2.5-4 percent range, with a 4 percent return exceeded since 1952 only in 1966 and 1972. The low earnings of the industry also can be attributed partially to deficit-producing passenger operations, only recently relieved by the transfer of these services to AMTRAK. The passenger-service deficit, as calcu- lated by the ICC yearly since 1963, is shown in the fourth and fifth columns of Table 3. The addition of these amounts to the net railway operating income figures in Table 2 would attenuate, but by no means reverse, the profit decline. In fact, in terms of total deficit (column 5), the passenger loss has been static or slightly dechning since the late 1960s. Net railway operating income, moreover, may actually overstate the earning power of the rail industry. In computing net railway operating income, the consumption of capital is measured by depreciation and amortization computed according to service lives specified by the ICC. I TABLE 2 Rate of Return by District United Eastern Year Southern States Districta District 1929 1939 1944 1947 5.30 2.56 4.70 3.44 6.03 3.14 4.37 3.02 4.27 2.77 1950 3.63 3.47 3.86 4.01 2.39 5.31 1952 1953 1954 4.28 3.76 4.16 4.19 3.28 1955 1956 1957 1958 1959 4.22 3.95 3.36 2.76 2.72 4.19 4.05 3.29 2.00 2.27 1960 1962 1963 1964 2.13 1.97 2.74 3.12 3.16 0.89 1.80 2.28 2.56 1965 1966 1967 1968 1969 3.69 3.90 2.46 2.44 2.36 3.32 3.55 1.53 1.27 1.10 4.16 4.45 3.86 3.79 4.17 3.87 4.03 2.75 1970 1.73 1971 1972 est. def. 2.47 2.95 del. 0.44 4.50 4.93 5.17 3.02 3.90 4.19 1951 1961 1.55 5.45 3.52 4.74 5.27 5.45 4.48 5.45 5.00 4.14 3.68 3.63 2.97 3.19 4.17 4.04 4.01 Western District 4.85 1.85 4.82 3.84 4.63 3.76 4.13 3.98 3.29 3.86 3.54 3.18 3.20 2.87 2.40 2.58 3.15 3.60 3.43 3.01 2.81 9ncludes Pocohontas r.gion. These accruals, in the view of many experts, value of the railroad plant. For one thing, understate the actual loss of the service lives accorded roadway and equipment are rather long, and much of rail investment has become economically obsolete before it has been fully depreciated in the accounts. In addition, because of the typically long service lives of rail assets and the continuing inflation in the U.S. economy, cost of replacement often substantially exceeds original cost. As in other industries, inclusion at replacement costs would increase depreciation accruals appreciably; in the railroad industry, with its exceptionally long accounting periods for depreciation however, the effective understatement based on historical costs may be considerably larger than for most other industries. I,' TABLE 3 Passenger Service Deficit, 1963-1971 (in millions) Passenger Year 1963 1964 1965 1966 1967 1968 1969 1970 1971 Service Revenues $1,107 1,085 1,042 1,018 878 686 638 586 407 Passenger Service Expenses, Taxes, and Rents Solely Relateda Total $1,116 1,103 $1,506 1,496 1,086 1,048 1,016 884 863 837 522 1,463 1,417 1,363 1,172 1,102 1,062 692 Passenger Service Deficit Solely Relateda $ Total 18 $399 410 44 421 31 400 485 486 464 477 285 9 138 198 225 252 115 lJnder accounting niles prescribed by the ICC, railroad operating revenues and expenses are divided into freight and passenger services so as to develop a net railway operating income for each service 'Solely related passenger service expenses" are expenses incurred directly irs the provision of passenger service and which could be avoided if passenger service were discontinued. "lotal passenger expenses" also include common and joint expenses that have been apportioned statistically between freight and passenger service and which could not necessarily be avoided if passenger service alone were discontinued, but which in all likelihood ssould be largely avoided if it were. [2] THE CHANGING MARKET FOR RAIL FREIGHT TRANSPORT Obviously, there can be little doubt that the U.S. railroad industry has not been highly profitable in recent decades. Of the many reasons advanced to explain this condiiion, unquestionably one of the most important is simply the changing character of the intercity freight market. Changing markets characterize the American economy, and the success of an industry depends in no small way on its ability to adapt the product it offers to changing demands. The freight market, like other markets, is subject to change and the problems of the railroad industry owe in large measure to its not being able to adapt to changing market conditions, in terms of both the types of services it provides and the manner in which it produces those services. As an economy matures, the evolution of consumption patterns and industrial activity causes the overall importance of materials to decline and the relative importance of various commodities to change. The volume of freight, particularly freight of a type suited to rail transport, normally fails to grow in proportion to the rate of economic growth. In gross terms, aggregate intercity freight tonnage has been growing only about 60 per- cent as fast as real GNP during the postwar period. The geographic distribution of economic activity also changes over time, partially in response to the changing composition of economic activity and partially in S 456 John R. Meyer and Alexander I. Morion response to the changing character of the transport system. This rekcation of industry and markets transforms the spatial pattern of freight flows, sometimes in ways that deprive the railroads of traffic and alter the densities of remaining rail traffic. The railroads participate, therefore, in what is inherently a slow-growth market and often an increasingly less profitable one as well. Bulk-Commodity Traffic It is useful to divide freight traffic into two categories: bulk commodities and manufactures. Bulk commodities are raw materials or "intermediate" (or "producers") goods in transit from raw-material origin to factory or between factories. Bulk commodities tend to be handled in flows rather than in discrete, packaged units, and they tend to move in large volumes. The commodities themselves are usually dense or "heavy" and of relatively low value per unit weight. They are typically not fragile, perishable nor particularly subject to depreciation in market value with time. The railroads' arch competitors for bulk-commodity traffic, on a cost basis, are the water carrier and the pipeline, though the relatively high-cost truck has made surprising inroads into even this traffic in certain special cases. The economic consumption of bulk commodities has not kept pace with the growth of national product. The postwar growth in total consumption of raw materials other than primary Construction materials (sand, gravel, and stone for direct use in construction), and air and water, has grown at an average rate o only 1 .9 percent per year during the postwar period, half or so the growth rate of about 4.0 percent in real GNP. Furthermore the long-rur trend is for the growth rate of raw material consumption to decline relative to the growth rate of real GNP. The explanation of this trend seems to lie in evolving consumption patterns and in advancing industrial technology For consumption increments to per capita income are increasingly devoted to the purchase of services such as health care, personal services, education, travel, communj-ation etc., rather entertainment than goods. Services typically require a lesser weight of physical materials per unit of output than manufacturing and construction Specifically, the total share of real GNP allocated to goods as opposed to services fell from 60.8 percent in 1945 to 48.1 percent in 1970. Even among manufactured goods, voted increasingly to goods relativelyincremental personal income is de- more labor- and materialjntensjve As personal incomes rise, consuners typically spend smaller fractions of their Incomes on such food, clothing, furniture, shelter, basic commodities as and fuel and larger fractions on such highlyfabrjcated products as than televisions, stereo sets, cameras, watches, The U.S. Railroad Industry in the Post-World War II Period: A Profile 457 jewelry, toiletries, and cosmetics, for which weight is much lower in relation to value. The 50 manufacturing industries among the 41 7 fourdigit SIC manufacturing industries exhibiting the highest rates of growth from 1958 to 1966 have freight input coefficients (as taken from the Department of Commerce's 1963 Input-Output table) averaging only one- half the weighted average freight input coefficient (0.02204) of all 417 manufacturing industries. In fact, as Table 4 reveals, only four among the 50 fastest-growing industries have freight input coefficients equal to or exceeding the 0.02204 average for all manufacturing. Another change in consumer expenditure patterns inhibiting material consumption is the tendency of consumers to substitute more expensive brands, models, and styles of individual products as their personal incomes rise. The weight of materials used in producing more expensive brands, models, and styles of commodities such as food, clothing, cars, furniture, appliances, and housewares rarely rises in proportion to the value of these products: in some products, weight does not rise with value at all.8 Progress in industrial technology also retards the growth of material usage. Economies in material usage are achieved in diverse ways, among them the substitution of lighter materials, development of synthetics, creation of more durable materials, miniaturization, improved product design, improved process design, new uses for by-products, waste and scrap, and improved inventory and distribution systems that reduce inventory "shrinkage." Even innovations less directly motivated by a desire to economize on materials and their transport nevertheless may have had such an effect. For example, nuclear energy replaces heavier fuels; radio, television, telephones, and computers are substitutes for "paper" forms of communication; airplanes and communication satellites require fewer ground structures and construction and utilize lighter equipment than their predecessors. The extent to which material-economizing innovations as well as the shifts to more highly fabricated products and to more expensive brands and styles have retarded the growth of material usage within individual industries is summarized in Table 5. The table contrasts the average annual growth of primary raw material inputs (in units of weight) for each of 16 major industries with the growth of each industry's output (as measured by the respective F.R.B. index of industrial production or some other index of industry output). In every case but one, output has grown more rapidly than the physical weight of inputs arid, by implication, the physical weight of outputs. Of course, volume and composition are not the only dimensions rele- varit to defining a traffic market. The spatial pattern also is important. Indeed, total ton-miles of traffic could increase even if volume declines, if miles traveled per physical unit rose by more than enough to offset the TABLE 4 [ow Freight Intensities of the Fifty Fastest-growing Manufacturing Industries - SIC Code 1929 3672 3674 3571 1914 2272 1951 3679 3339 3831 3392 3742 3537 3399 1961 3841 3861 3953 3541 3663 3651 2262 2519 3622 3357 I ndustr,, Ammunition, N.E.C. Cathode ray picture tubes Semiconductors Computing and related mach. Guns, mortars, ordnance N.E.C. 5.80 5.10 4.49 4.38 4.15 0.00611 0.00764 0.00739 0.00383 0.00551 Tufted carpets & rugs Small arms Electronic components Primary non-ferrous metals, N.E.C. Optical instruments & lenses 3.80 3.46 3.35 3.33 0.01256 0.01802 0.008 14 0.0086 1 3.23 0.00630 Non-ferrous forgings Railroad & Street cars Industrial trucks and tractors Primary metal industries, N.E.C. Small arms ammunition 3.16 0.01055 0.01842 Surgical and medical instruments Photographic equipment Marketing devices Metal-cutting machine tools Radio, TV, communications equipment Radio and TV receiving sets Textile finishing plants, synthetic Household furniture. N.E.C. Industrial controls Non-ferrous wire drawing 3693 2282 3799 3499 3079 Plastic products, N.E.C. 2295 3545 3479 3715 3565 Coated fabric, not rubberized Machine tool accessories Metal coating, engraving, etc. Truck trailers Industrial patterns 3536 Hoists, cranes & monorails Trailer coaches 3791 Growth Ratio (value of 1966 Direct Freight shiprrients Input value of 1958 Coefficient, shipments) i 963 X-ray & therapeutic apparatus Throwing and winding mills Transportation equip., N.E.C. Fabricated metal prod., N.E.C. 3.02 2.94 2.8.3 0.0148 1 2.79 0.00777 0.01274 2.76 2.73 0.01143 2.71 2.69 2.65 0.01317 0.00787 0.00373 2.64 0.01177 0.0 1005 2.61 0.015 58 2.55 2.54 0.02669 2.53 0.01393 2.53 0.006 16 2.51 2.50 2.50 2.49 0.01875 0.02679 0.00945 0.01368 2.41 2.41 0.0 1035 2.39 2.38 2.37 2.37 2.36 0. 00740 0.00598 0.01071 0.0164 1 0.00359 0.01127 0.01865 TABLE 4 (concluded) SIC Code Industry Growth Ratio (vdue of 1966 Direct Freight shipments Input value of 1958 Coefficient, shipments) 1963 2256 Secondary non-ferrous metals Copper roIling and drawing Knit fabric mills 2.36 2.35 2.35 0.05940 0.02630 3713 358X 2327 2844 3542 Truck and bus bodies Service industry machines, N.E.C. Separate trousers Toilet preparations Metal-forming machine tools 2.33 2.33 2.30 2.29 2.29 3559 3572 3569 3451 3694 Special industry machinery, N.E.C. Typewriters General industry machinery, N.E.C. Screw machine products Engine electrical equipment 2.27 2.24 2.23 2.23 0.00927 2.21 0.0 1068 1.65 0.02204 3341 3351 All manufacturing SOURCE: 0.009 19 0.0 1420 0.01 302 0.00511 0.01295 0.00894 0.0082 1 0.00684 0.01246 U.S. Deparrrnenr of Commerce, !ndosry Profiles, i958-r 966 (L'S. Governnsenr Prinring Office, 1968); and U.S. Deparrment olCornmerce, Input-Output Structure of the U.S. Economy. (963 (U.S. Government Printing Office, 1963). tonnage drop. The geographic organization of economic activity does, of course, adapt itself to changes in the economy. To start, manufacturing is becoming more market-oriented, for many reasons. For example, manufacturers usually ship bulk commodities rather than finished goods because of the lower freight rates for bulk corrlfnodities.bo Similarly, the shift in competitive emphasis from price to style and service encourages producers to locate as close to markets as possible. The concentration of people, and therefore of demand, in large metropolitan areas makes it easier to achieve economies of scale in manufacturing in more and more regions, thus making decentralization and a market-oriented location increasingly feasible. Another factor tending to lengthen hauls of bulk commodities is increasing geographic specialization in the production of materials. The various regions of the nation are not uniformly endowed with natural resources. As freight rates decline, it becomes economical for a market to draw on ever more distant raw material sources. In particular, the decline in freight rates with distance has encouraged the use of more distant, low-cost raw material sources. The United States has thereby progressed from an economy in which local areas were nearly self-sufficient in producing their own raw material needs to one of regional self-sufficiency (in which there was TABLE 5 Growth of Primary Raw Material Inputs (in tons) Compared with Growth of Output (as measured by F.R.B. indexes of industrial production), 1947-1 967, by industry Average Annual Sector or Industry of Primary Raw Material Inputs, 1947-1967 (in order of declining weight of primary raw material inputs) (%) New construction 22 Energy 1.9 Manufacturing Petroleum refining Clay, glass, & stone products Food products Chemicals & products Priniary ferrous metals Primary non-ferrous metals Fabricated metal products, machinery, transportation equipment, & instruments & related products Lumber & products Paper & products Rubber & plastics products Textiles Apparel 2.2 3:3 2.9 2. 1 4.1 3.6 2.8 4.5 3.8 3.7 2.5 8.3 1.0 1.5 3.4 2.1 5.5', 1.7 Alexancr L Morton, 'Freight Demand," 1947i 967 (F.R.B. Industrial Prod. Index) (%) 2.8 0.7 2.7 7.6 Leather SOURCE: Average Annual Growth in Weight Growth in Output, 1.4 4.6 6.4 2.6 1.9 3.5 --1.6 0.7 Ph.D. dissertatiwi Harvard Univeity (unpubIish Chapter 5. The index used us the Composite Index of (Ten) Construction Materials, published by the Construction and Building Materials Divison of the Department of Commerce. The ten materials are weighted ri relation value, so that the fapid growth of sand, gravel, and stone to otherwise conSumption affecix the index less sharply than rsirnple average of the F.R.B. production indexes of the (Our categories. considerable trade within regions, but regions themselves tended to be self-sufficient) and presently to one of considerable regional iriterdependence (in which many raw materials are produced in whichever areas of the Country have a comparative advantage). Indeed, many important commodities, including Citrus and other fresh meat, lumber, iron ore, coal, and petroleum fruits, and vegetables, grain, are supplied throughout the nation from the one or two regions in which they are produced most cheaply. The exhaustjor of natural resources close to historical population Centers also causes markets to turn to more distant Sources of supply, thereby The U.S. Railroad Industry in the Post-World War It Period: A Prutile 461 lengthening hauls of bulk commodities. This process of exhaustion and discovery of new sources at more remote locations is or has been at work in the case of coal, petroleum, iron ore, lumber, and soil fertility. Increasing geographic specialization and the exhaustion of older sources of supply have generally pushed production of raw materials into the west and the south. Most of the traffic growth arising from the lengthening of hauls has, therefore, occurred in these regions. For those commodities already produced in but one or two areas of the U.S., regional specialization already is well developed, implying that additional lengthening of domestic hauls of these commodities is improbable. Further geographic specialization, if it occurs, is more likely to be on an international scale. Although the effect of expanding international trade on the demand for domestic intercity freight will depend on the types of goods and services exchanged, it may be generally negative. Exports of bulk commodities have only to reach ihe nearest port. As for imports, to the extent that the population is increasingly concentrated in metropolitan belts extending along the Atlantic, Pacific, and Gulf coasts, and around the Great Lakes, all accessible by ship, imports of bulk commodities and manufactured goods received in trade can be delivered to an increasing proportion of American markets without a long domestic intercity overland haul. Although intercity hauls of commodities generally seem to be longer, intercity movements of some commodities can become shorter, or even be eliminated altogether, for various reasons. For example, in some industrial sectors certain materials can be substituted for others and these substituted materials may have very different accessibility characteristicssome may originate at great distances whereas others may be available nearby. Thus, the construction industry can choose among concrete, metals, lumber, and synthetics as structural materials, as insulating materials, and as facing or surfacing materials, and these substitutions can be sensitive to differential freight costs, just as they are sensitive to cost differentials in general. Similarly, electric utilities may switch fuels, so that differences in cost, including freight charges, may alter the choice among, say, nuclear, oil, pipelined natural gas, rail transported coal, or hydroelectric or mine mouth generation (with subsequent high voltage transmission). Obviously, not only the total distance and volume of transport involved, but the modal choice will be sensitive to which of these alternatives is selected. Recycling of used materials presents industry with another opportunity for shortening or even eliminating some intercity freight hauls, for recycled materials substitute for virgin materials that typically must be brought to the market over long distances. Recycling is presently limited to minor amounts of metals, paper, rubber, and industrial chemicals, although it appears probable that the proportions of materials recycled will increase S 462 John R. Meyer and Alexander 1.. Mortor and that recycling will be extended to other commodities, such as heat energy, as the technology is developed. Rising prices oi virgin materials as resources are depleted and increasing concern over the environmental impact of waste disposal also can be expected to encourage recycling. Good estimates of the net effect of all these conflicting influences on the length of haul for bulk commodities during the postwar years are difficult to ascertain. The average length of haul for all rail traffic apparently has grown from approximately 411 miles in 1947 to 488 miles in 1967, or at an annual rate of 0.9 percent per year.'1 This average, of course, represents both movements of bulk and manufactured commodities. For water carriers and pipelines, which are more heavily specialized in the movement of bulk commodities than railroads, annual growth rates in average length of haul in postwar years have been 0.2 percent and 0.9 percent, respectively.2 It is difficult to believe that the length of haul of rail bulk commodities has increased at a rate any more rapid than that for pipelines during these postwar years. This would suggest that the overall rate of growth in length of rail haul of 0.9 percent per year, which is essentially identical with that for pipelines, would be a reasonable estimate of the actual rate of increase in bulk-commodity trip lengths over the postwar period. Assuming this to be the case, this 0.9 percent growth per year in length of haul would constitute almost one-half of the remaining gap between real U.S. GNP growth, approximately 4.0 percent per year during the postwar years, and the average 1.9 percent per year growth in total physical tonnage volume. in short, lengthening of hauls has probably increased the actual rate of growth in total rail bulk ton-miles of traffic during the postwar years by approximately 50 percent, but this is still not enough to bring the rate of increase in this market up to parity with the rate of increase in real GNP. Traffic in Manufactures The other major category of freight traffic, manufactures consists primarily of finished goods, or parts, moving from factory to assembly plant, warehouse, or store. Manufactures typically move in packages or discrete units; shipments tend to be of lower volume, smaller and less regular, than bulk-commodity flows. The goods themselves tend to be of lower density (i.e., lighter) and of higher value per unit of weight. They frequently are fragile or perishable and subject to loss of market value as a function ot time. This traffic accounts for 25-40 percent of railroad ton-miles (depending on definition) and a still larger proportion of revenues and profits. Obviously, the weight of manufactured goods must bear some reasonably constant relation to the weight of raw material inputs. Accordingly, the weight of traffic in manufactures almost certainly has been growing more slowly than real GNP. However, it is significant that aggregate intercity The U.S. Railroad Industry in the Post-World War II Period: A Profile 463 freight tonnage in both bulk commodities and manufactures has been growing at a rate about 60 percent as fast as real GNP during the postwar period whereas, as already mentioned, the rate of increase in bulk commodities alone has been growing only about 50 percent as quickly as real GNP. The discrepancy of 10 percent in growth rates between bulk commodity and total commodity tonnage suggests some increased activity in manufactured goods traffic. The primary cause of this relatively faster growth in manufactured goods traffic tonnage would appear to be an increase n the number of shipments. For example, as per capita incomes rise, there is a shift in consumer expenditures toward products that pass through a greater number of stages of fabrication.13 Increasing specialization or 'division of labor" in manufacturing industry also contributes to a rising number of interplant shiprnents. Hence, the total tonnage of manufactured goods shipped may rise more rapidly than bulk commodity tonnage. The proliferation of distinct brands, models, and styles of manufactures may also reduce the size of individual shipments and require a more elaborate system of distribution. It is not so much in tons or ton-miles, however, as in freight revenues that the faster growth of traffic in manufactures than in bulk can be discerned (i.e., in greater revenue per ton-mile). Aggregate intercity freight revenues have been growing about 86 percent as fast as current dollar GNP and the relationship would not seem to be significantly altered if one uses price deflated or real values (since the deflators for freight revenues and GNP are not too dissimilar). Freight revenues from bulk commodities have apparently grown roughly in proportion to bulk-commodity tonnages, economies in bulk handling apparently offsetting the cost of slightly longer hauIs.' Accordingly, the greater growth in total revenues than in total tonnage has derived chiefly from traffic in manufactures. An increase in the quality of freight service used is the major explanation of these differential growth rates between tonnage and revenues. A funda- mental change in the postwar freight market is the increasingly higher standards of service that shipments of manufactures require. This change parallelsand, in large part, derives froma shift in most markets toward higher-quality products and services. As the economy shifts to production of more highly fabricated goods and more expensive brands, models, and styles, the value of manufactures per unit of weight tends to rise because of the relatively greater inputs of labor and capital. As the amount and cost of working capital tied up in goods rises, there is a tendency to opt for speedier, more reliable deliveries as a way of controlling logistical costs. Growing sophistication among shippers concerning the trade-offs between transport and other components of the total distribution bill also is partly responsible for the gradual shift to higher-quality freight service.'5 Highly fabricated and expensive goods are inclined to be fragile, perish- 464 John R. Meyer an(I Alexander L Morton able, or otherwise damage-prone prompting the USe of premium service. Greater stress on styling erodes rap!dly the market value transpo,-t of many products, which also stimulates demand for speedy, reliable delivery Increasingly, products compete in the market on the basis of quality and service as well as price. The reliability, speed, and convenience of deliveries is part o1 the quality of service for which a buyer or shipper looks. Unfortunately for the railroads, they have not been well meet this demand for higher service standards. The easiest equipped to way of doct, menting this point is to look at the trends in intercity freight shares by the different carriers, shown in Table 6. The most important statistics in Table 6 for evaluating trends in manufactured-goods traffic are the relative performance of rail and truck shares, since the truck is unquestionably the major competitor of rail for this traffic. Since the end of World War II, truck ton-miles have increased fivefold and the truck share of the total market has more than tripled; railroad ton-miles, on the other hand, have increased only slightly, if at all, and the rail share of the total market has declined from roughly two-thirds to a little less than 40 percent. Without much question, the superior growth performance of the trucking industry is in large part attributable to that industry's better performance in meeting the increasingly higher service requirements for traffic. manulacturedgoods Changes in the spatial patterns of traffic and manufacturing also may have handicapped the railroads in competing with trucks. Probably the most important change is the urbanization of the American population The rural population of the U.S. has declined as a percentage of the total Population from 85 percent in 1850 to 60 percent in 1900, 36 percent in 1950, and 25 percent at present. A rural population, geographically disbursed about its manufacturing and distribution centers, requires intercity hauls for distribution of its consumer goods. Much of the rural rail network was constructed to provide both freight and tion to rural passenger transportacommunities that, at the time, had no suitable alternative means of transport. The delivery of consumer goods to a rural population supplied the railroads with traffic that was doubly valuable (1) because manufacture tend to be relatively high-rated (i.e., profitable) traffic; and (2) because manufactures can be while bulk commodities are being back-hauled into rural communities migrated to cities, this flow of moved out. As the rural population has manufactures has diminished relatively, causing much of the rural rail network to become superfluous In essence, urbanization has eliminated commodity flows out of ruralany semblance of balance between bulk areas and manufactured goods back, a io-way traffic for which the all-purpose rail boxcar was uniquely well suited. The growth of large urban markets also encourages the decentralization 465 ton The U.S. Railroad Industry in the Post-World War II Period: A Prouile ort and market orientation of manufacturing, a factor that tends to shorten or do away entirely with some intercity hauls of manufactures. Indeed, given that the strength of these demographic trends, it is somewhat sur)riSing total manufactured-goods traffic has continued to grow as much as it has. manufacAt any rate, the minimum efficient scale of production for many turing processes has not grown so rapidly as the size of the market in most metropolitan areas."' Hence, an increasing proportion of the total market for products is concentrated in cities of a size sufficient to support their own factories in an increasing portion of the manufacturing spectrum. That is, metropolitan areas are becoming increasingly self-sufficient in their local manufacturing capacities.17 There is, accordingly, a tendency toward a diminished flow of manufactured goods among cities, ceteris paribus. The traffic growth lost as a result of such "import substitution" is again a fly fry. and de)kS. to Cuthe Ic 6 ora jo r uck rket inhas hout ry is ti ng ods may the tion. total nt in cally nterrail ortaative I .tion ause and nities has ively, ence, bulk ck, a well atiofl relatively high-rated traffic. The flows also tend to be of relatively highdensity, and therefore are less costly to move on a unit basis, since the number of urban origins and traffic is concentrated among a limited destinations.'8 Finally, to the extent that each metropolitan area must produce something in trade, these intercity flows tend to be somewhat balanced as to back-haul. This decentralization or regionalization (or megapolitization) of manufacturing, in concert with the migration of population to the west and the south, also is having an effect on the interregional balance of traffic in manufactures. Even after the discovery and development of abundant natural resources in the west and during the early westward migration of markets, manufacturing remained highly concentrated in the "industrialized northeast." This generated long hauls of raw materials into the northeast and long hauls of manufactures back to the west and the south. It could only be a matter of time until some manufacturing capacity migrated in order to abbreviate this roundabout traffic. The dispersion of manufacturing away from the northeast thus contributes to lower traffic volumes and the redundance of fixed rail plant in the northeast. Concurrent with the migration from rural to urban areas has been a migration of population and industry from the central cities to the suburbs. The population residing in the "outside-central-city" areas of the Standard Metropolitan Statistical Areas (SMSAs) grew at a rate half again as great as the growth rate of the central cities between 1930 and 1 950, three times as great between 1950 and 1960 and, it is estimated, will grow six times as rapidly between 1965 and 1 975,19 Commerce and industry are suburbanizing at about the same rate as the population.2° Suburbanization seems to be generally adverse to the fortunes of the rail industry. Railroads were built at about the same time as cities were expanding or forming, and cities in most cases either grew up around rail facilities or else rail terminals were built near the industrial cores of the 9 379 482 645 735 747 691 602 665 647 535 597 655 623 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1955 1956 1957 655 626 631 614 557 339 1939 1954 Amount Rail 49,5 48.4 46.9 56.2 55.6 54.4 51.0 49. 67.2 66.6 65.3 61.9 58.3 61.3 62.4 69.5 71.3 68.7 62.3 % 223 249 254 217 213 173 188 195 67 82 102 116 127 60 57 58 81 62 53 Amount Truck 17.5 18.4 19.0 16.3 16.0 17.0 18.0 19.0 13.8 11.1 10.0 9.1 6.5 10.0 10.5 6.5 5.5 5.3 9.7 % Intercity Freight by Carriers (billions of ton-miles) Year TABLE 6 Oil 203 230 223 129 152 158 170 179 127 96 105 120 115 59 68 75 98 133 56 line Amount PipeLakes 91 117 111 119 127 14.1 15.9 15,9 17.0 16.7 105 112 120 96 112 119 98 113 96 114 122 115 119 76 Amount 12.9 13.8 12.1 12.4 10.6 10.3 11.5 12.5 9.5 12.2 8.1 9.5 8.8 10.3 % Great 109 115 ' 98 9.3 52 62 64 75 30 28 35 43 42 27 26 26 22 20 Amount 8.2 8.8 8.1 10.5 10.2 9.2 10.6 11.0 10.6 11.0 11.4 10.7 ifl 11.2 13.1 15.5 14.8 14.0 % Rivers and Canals 4.1 7.7 8.0 8.6 5.3 5.6 6.2 7.4 4.9 4.6 0.49 0.58 0.68 0.30 0.34 0.34 0.37 0.38 0.11 0.15 0.20 3.4 3.1 0.09 0.08 0.02 0.02 0.04 0.05 0.07 0.01 Air Amount 2.9 2.8 2.5 3.6 3.5 2.8 3.7 % 0.04 0.04 0.05 0.03 0.03 0.03 0.03 0.03 1,274 1,356 1,336 1,063 1,177 1,145 1,203 1,123 1,028 904 1,019 1,045 917 1,088 0.01 0.01 0.01 0.01 0.02 1,031 0.01 618 772 928 544 Total 0.00 0.00 0.00 0.00 0.00 % I 412 422 39.7 38.7 757 774' 758 744 1970 1971(p) SOURCE: NOTE: 389 396 404 229 233 238 253 269 306 333 21.8 22.7 22.5 21.9 21.8 22.0 21.5 21.3 21.3 21.9 22.3 23.1 444 20.5 21.3 21.7 19.1 18.7 17.3 17.4 17.4 17.4 17.8 17.4 17.7 16.7 431 411 361 391 227 23.1 23.1 211 21.7 223 21.1 19.0 114(79) 104 110(76) 116(81) 107(75) 112(75) 115(83) 99 87(67) 90(66) 95(68) 106(73) 80 8U 117 1,314 1,310 1,371 1,453 1,543 0.07 0.013 0.09 0.09 0.10 0.89 1.01 1.30 1.30 1.50 9.2 9.4 9.7 9.6 9.3 1,216 1,286 0.01, 0.05 0.70 0.80 1,933 1,923 0.15 0.18 3.40 3.50 9.4 9.9 9.7 9.9 10.6 10.7 205(1 56) 205 6.0 5.4 6.1 6.1 6.1 1,638 1,747 1,765 1,838 1,895 0.12 0.13 0.15 0.16 0.17 1.91 2.25 2.59 2.90 3.20 9.3 152(110) 164(117) 174(128) 179(139) 188(144) 9.1 9.0 6.7 6.6 6.6 6.5 6.9 121 7.5 6.6 123(84) 133(90) 139(94) 144(102) 109 117 6.6 6.2 8.8 Transportation Facts and Trends, july 1972. lntercityfreiht by carriers includes both for-hire and private carriers; p) indicates TAA preliminary estmate. Effective 1969; no Ioner includes mail and express. 751 731 381 43.3 43.0 41.4 41.2 40.8 709 1965 1966 1967 1968 1969 1962 1963 1964 359 285 296 309 336 356 43.5 43.8 43.3 43.2 44.1 579 570 600 629 666 1960 1961 256 279 254 46.0 45.3 46.9 559 582 626 1958 1959 1957 John R. Meyer and Alexander L. Morton newly-forming cities. As long as cities remained dense and clustered around the rail facilities on which they were so depen(l(,lt traffic readily moved by rail.2 Rail lines have not been extended in most Cjti('S to serve emerging suburban areas as completely and efficiently as they Serve the central business district and, indeed, probably could not be for technological reasons. Suburbanization, therefore, tends to carry consum ers, warehouses, and factories away from rail service.22 Furthermore in choosing among possible suburban locations, factories, warehouses and shopping centers are sometimes more concerned today with locating along highways for ease of access by employees and customers than they are with locating along rail lines. Not only does suburbanization reduce the Proportion of shipments with origins .and destinations having ready or efficient access to rail, it also transforms the pattern of movements to a pattern for which tile rail network and conventional train operations are not well suited. Formerly the distribution pattern for manufactured goods tended to be radial, outward from urban manufacturing cities to satellite cities and towns, thus Paralleling the rail network. The present trend is toward movements of manufactured goods that originate at one suburban point and terminate at other suburban points scattered about the metropolitan area. This emerging distribution pattern is less radial, more dispersed and 'random," with an increasing number of shipments moving over trans-suburban routes not paralleled by rail lines and lacking the density for conventional train Operations In sum, the basic trends in manufacturedgoods traffic have not been so adverse as those for bulk commodities Certainly, the aggregate from manufactures traffic have almost kept revenues realized in pace vith the growth in GNP. However, from the standpoint of the railroads, this relatively favorable performance of the total market for manufactured goods has not been particularly beneficial, as most of this traffic growth has gone to other carriers, and especially to trucks. Trucks apparently were better adapted to rendering the higherquajjty service increasingly required by manufactured goods, and the spatial distribution of activity has trencjed away from situations and sites (i.e., origins and destinations) that once were well served by the railroad route structure. 131 RAIL PRODUCTIVITY Traditional Measur GROWrH of Rail Productivity Productivity growth measures the efficiency with which an industry improvement over the years in the converts its factors of production into a ri t, ) r The U.S. Railroad Industry in the Post-','Vorld War II Period: A Profile 469 salable service or product. In an environment in which competing iridustries are achieving significant productivity gains, an industry whose productivity lags normally finds it difficult to simultaneously match wage increases in other industries, keep its prices stable to fend off competition from other products and services, and maintain a satisfactory rate of profit. For example, when productivity lags, market-enforced wage increases must be translated into higher prices, which gradually cause demand to cease growing and perhaps eventually to taper off. At the same time, lagging productivity and falling output put pressure on profits and the ability of any industry to raise capital for renewal and modernization. Accordingly, industry growth in output and productivity generally are positively correlated. It is at least somewhat surprising, therefore, given the static character of output growth in the railroad industry, that published studies of industry productivity growth rates suggest that the railroad industry has done quite well by usual measures of productivity gain. Specifically, "net ton-miles and passenger-miles per man-hour," the index (attributable to the U.S. Department of Labor's Bureau of Labor Statistics, or BLS) frequently used to measure productivity in the transportation industries, has grown at an average rate of 5.2 percent per year in the rail industry during the period 1 947i 970 (and 6.0 percent per year between 1957 and 1970). During the same period, the productivity of the private economy as a whole, measured as real net output per man-hour of labor, has grown at an average rate of only 3 percent. According to these studies, then, the railroad industry has generated productivity gains at a rate 70 percent greater than the rest of the economy.23 In addition, for the raIroads' major competitors, output per unit of labor input rose only 3.1 percent per year in intercity motor trucking and 0.7 percent per year in water transportation vs. 5.8 percent a year in railroading during the period 1948-1 966.24 The question thus arises, how have the railroads seemingly violated the usual rule relating good productivity performance to general industry growth and prosperity? At least one explanation may be that net ton-miles per man-hour is an incomplete measure of overall rail productivity, at least for some purposes. To begin, rail labor inputs have declined much more rapidly than capital inputs. In most industries capital has been substituted for labor over the last few decades but this substitution has been especially pronounced in the rail industry. The capital-labor ratio increased at an average rate of 2.6 percent per year in the private domestic economy during the period 1948_1966.?S In the rail industry the capital-labor ratio has increased at an average annual rate of about 4.7 percent during the same period. Like capital inputs, inputs of materials and services (or so-called "intermediate" inputs) have tended to decline less rapidly than labor inputs in the rail industry. Specialization is an important means of improving pro- S 470 John R. Aleyer t)(1 on ductivity, and the railroad industry has used this device extensil Not only has it increasingly employed relatively more outside labor servic but it also has turned increasingly to purchasing the services of capital assets owned by others as an alternative to owning the particularly evident in rolling stock. Although the assets itself This k number of railroad owned freight cars has been decreasing throughout the POStwar years the number of cars leased by the railroads and the number of cars Owned by car companies and shippers have been increasing until ifl 1971 the latter accounted for roughly 18 percent of all freight cars.2 In 1955 railroads built over half as many freight cars as the car-building industry, but by 1971 they built only 15 percent as many. Recorded man-hours may also understate the growth of iflp(j5 needed to produce current output while maintaining rail plant at given standards7 Financial stringencies have prompted some rail companies to economize on labor and other inputs by deferring maintenance of their properties and by allowing the average quality of the rail plant to deteriorate This "reduced-level" of maintenance may be wise and proper in cases in which there is an excess of rail plant that will never need to be replaced. generally, deferring maintenance more may simply postpone e)(penses associated with producing present output, thereby causing present inpu to be understated and productivity gains to be overstated 28 Ton-mileage might also overstate the output growth of the rail indust This could occur because, among other reasons, there has been a shift in the composition of traffic toward that which is inherently less costly to produce and less valuable in the market place. The most impo,lant instance of this is the decline of passenger service. Since passenger5 are several times more costly to produce than revenue freight ton-rn iles the precipitous decline of Passenger traffic relative to freight traffic has caused a reduction in average cost per unit of the composite output usc-d by the B[S and others. In addition freight ton-mileage has come from virtually all the growth in aggregate the growing length of freight hauls. As pointed out previously, average distance per ton increased in recent years_-.indeed, by approximately 24 percent between 1947 and 1971. The marginal factor input requjrem of longer hauls is comparatively light, that growth of ton-mileage so stemming from longer hauls is a less costly increment to output than output were measured asone stemming from added shipments. Thus, if revenue tonnage rather than ton-mileage, measured Productivity growth between 1947 and 1971 24 percent less. would be roughly As a measure of output ton-mileage also does tion in quality of service that may be overtakingnot reflect any deterioraextent that a lower quality some railroads. To the needed to produce a given of service is less costly to perform, the inputs ton-mileage are red,;ced, causing conventional I ri ot 5, al is e y er 'S y to e d Is h e 5- 0 n 0 S S e S The U.S. Railroad Industry in the Post-World War II Period: A Profile 471 productivity measures to rise. Standards of rail service may have fallen in a number of ways. Shippers, it is alleged, must wait longer for empty cars for loading, and more often the cars on arrival are dirty or otherwise unsuitable for loading. Speed of delivery has shown little if any improvement during the postwar period and may even have declined; the average freight-car cycle time (the time elapsed between successive loadings of a single car) has increased from 16.6 days in 1947 to 25.5 days in 1971, an increase substantially in excess of the 24 percent rise in length of haul. The incidence of loss and damage, another aggravation to shippers, has shown a fairly steady upward trend. As a rough indicator, loss and damage claims paid have risen from $98 million, or 1.1 percent of freight revenues, in 1955 to $235 million, or 2.0 percent of revenues, in 1971. Sensitivity Tests of Productivity Measures The standard measures of rail productivity can be modified or tested for sensitivity to the various factors cited above. The first or most obvious of these would be to adjust for the changing character of rail output. The unit of rail output used by the BLS is a combined total of ton-miles and passenger-miles in which a passenger-mile is said to be the equivalent of roughly two freight ton-miles, corresponding to the fact that average rail passenger revenue per passenger-mile is approximately twice the average rail freight revenue per ton-mile. However, as stated in section 1, throughout the postwar period passenger service has been a cross-subsidized and deficit-producing operation, so the relative revenues of passenger and freight operations fail to reflect their relative costs. Using standard ICC formulas for apportioning operating expenses between freight service and passenger service, passenger operating expenses per passenger-mile have been estimated to be from five to nine times as great as freight operating expenses per tonmile.29 Counti rig passenger-miles as the equivalent of five rather than two ton-miles, total rail traffic measured in ton-miles would be found to have declined at an average of 0.3 percent a year between 1 947 e and 1970 rather than rising at an average rate of 0.2 percent as the BLS statistics indicate. This, in turn, would reduce the measured labor produc- 0 tivity gain by one-half of one percentage point per year from the 5.2 y if percent per year calculated by the BLS to 4.7 percent per year. Of course, y e S if one were to count a passenger-mile as equivalent to seven or eight freight ton-miles, the simple labor productivity gain using the BLS index would be further reduced. As noted, the use of ton-miles as the unit of freight output may also overstate the growth of rail output, for the shifts toward lower-value, bulk commodities, toward larger shipments and longer hauls, all are shifts toward ton-miles that are inherently less costly to produce. Assuming that 472 john R. Meyer and Alexander 1. Morron the relative costs of different types of freight shipments are roughly prop tional to their respective revenues ti.e., applying the BLS rule for weighting passenger and freight service to weighting different types of freight traffic which would seem more defensible for freight service since freight mark- ups are probably more uniform), an estimate of the actual increase in freight output may be obtained by deflating total freight revenues by an index of freight rates (say the RI-i freight rate index computed by the ICC); by so doing, one can at least partially abstract from systematic changes in the composition of freight traffic with respect to commodity, size of individual shipments, and length of hauls. Freight manner would have risen only 0.2 percent per output measured in this year between 1947 and 1970, one-half a percentage point less than the 0.7 percent per year increase in revenue ton-miles.30 If translated into an adjustment of the BLS productivity statistics, this would reduce the measured productivity growth in the rail industry by another ½ of 1 percent per year. Together, these changs in the output indexone to weight two in accord with its contribution to operating passenger service more ready expenses, the other to account for the changing character of freight traffic.woukl reduce the measured rate of growth in labor productivity by a full percentage point. To the extent that the rail industry, like most other industries, has substituted capital for labor in the production process, productivity may also be better measured for many purposes by endeavoring to combined or multiple factor productivity construct indices. Specifically, Kendrick, in a study of productivity trends in the American economy, defines what calls "total factor productivity" which he takes into account increases capital as well as labor inputs.3' in Indexes of labor and capital "on the basis of the labor and are weighted together in the Kendrick study property shares of national income ing in the railroads originatas estimated by the Office of Business The property share of capital Economics."32 in national income, however, profits, and inasmuch as the corresponds to rate of profit in the highly regulated industry is relatively low, the railroad level of capital inputs could understated by the Kendrick proportionate be procedure. As pointed rate of return on net investment out in Section 1, the (original cost less depreciation) transportation property has averaged in rail only about 3 percent during the postwar period. This is only half the 6 percent traditionally taken as a desirable return that the l.CC has standard for the rail industry and one-third the 9 percent only to 10 percent or more after-tax return that is perhaps more "normal" throughout non-regulated It could be argued, therefore, private industry. that productivity industry might be more measures for the rail nearly comparable capital inputs were accorded with unregulated sectors if them in the Kendrick study. two to five times the weight that is accorded (On the other hand, of course, Kendrick can n g The U.S. Railroad Industry in the Post-World War II Period: A Profile 473 argue that all of his indices are constructed using the same definitions in all industries SO that inter-industry comparability is thereby improved.) Ken- drick affords capital inputs a weighting of roughly 10 percent and labor inputs a weighting of 90 percent during the period 1950 to 1966. The effect of a heavier weighting of capital inputs on the measured growth of Kendrick's total factor productivity would be to reduce it by three-tenths of a percentage point for each doubling of measured capital inputs; thus, with a 75-25 weighting instead of Kendrick's 90-10, total productivity growth from 1948 to 1966 would drop from Kendrick's estimate of 5.2 percent to 4.7 percent.33 Alternative Measures of Rail Productivity Total rail output, for purposes of calculating the growth of factor productivity, has been, as we have seen, traditionally a weighted combination of revenue freight ton-miles and revenue passenger-miles. However, the growth of freight traffic might be better represented not as the growth of physical ton-mileage, for reasons advanced earlier, but as the growth of rail freight revenues measured in constant or deflated dollars, There is also the question of how to combine passenger traffic with freight traffic in order to arrive at a measure of total rail output. Using the convention that a passenger-mile is equivalent to five revenue freight ton-miles (to reflect the relative cost of producing passenger-miles vs. ton-miles), railroad output for the years 1947 and 1970 could be measured as follows: Freight servicein ton-miles (billions) measured in constant rail freight dollars Passenger service in equivalent ton-milespassenger-miles (billions) x 5 Total output in ton-mile equivalents (billions) Total output index (1947 = 100) 1947 1970 655 689 230 885 100 54 743 84.0 Total inputs to the rail industry consist of labor, capital, and purchased materials and supplies. For productivity measures, each input, like each output, should be measured in real or constant dollar terms or physical units or as close thereto as possible. Accordingly, labor input might be measured as customary by man-hours worked, including straight time and overtime.35 For capital inputs the measurement problem is considerably more complex. Fortunately, the Bureau of Accounts of the ICC in 1937 reevaluated all the transportation property (other than land and landrights) of the railroads; they concluded that the cost of reproducing the rail system Ion TABLE 7 Capital Stock, Owned by Railroads and Employed in Producing Rail Output, Reproduction Cost Less Depreciation 9l Constant ad lie for he Current Dollars Year 1957-1959 Dollars 1947 1948 1949 51.7 51.7 52.3 34.0 37.6 950 52.5 1951 53.1 1952 1953 1954 53.5 53.7 53.4 38.8 41.7 43.3 45.0 45,2 1955 1956 1957 1958 1959 53.2 53.2 53.3 52.8 52.4 45.8 49.4 1 960 52.1 1961 1962 1963 1964 51.5 51.0 50.8 50.9 53.4 52.5 52.0 51.6 52.2 1965 1966 51.3 52.0 1 967 52.1 51.9 38.1 37 .5 ly ue he ge ck ue ry all C al phe of ce or 55 to d e 1968 1969 1970 SOURCES: is rst 52.1 52.9 53.5 51.9 52.9 54,5 55.9 58.8 61.8 51.8 65.3 Kendrick, Productivity Trends in U.S. Transportat ion lndustties. (prepared for the Office of the Under-Secretary of Transportation, U.S. Department of Commerce, January 1966), pp. 8, 31 -.32; and U.S. Department of Commerce, Survey of Current ffusiness, June 1956 and March issue of 1958 and subsequent years. Of e d ly es d excess of gross capital expenditures above depreciation in the financial accounts of the railroads may not reflect net physical additions to the capital stock so much as the rise in the cost of replacing assets over their original cost.36 ly The preceding estimates of the rail-owned capital stock exclude land ar ss and landrights. It is estimated that railroads occupy roughly 3 million acres of rural land and 300,000 acres of urban land, only slightly changed between 1947 and 1970. Assigning value to these lands at the rate of $100 per in rural acre and $10,000 per urban acre, roughly corresponding to the C average value of rural and urban acreage in 1957-1959, the total value of Is John R. \le,t'r ,ni(i \l \1flhICr L the railroad investment in land woul(J be estimated at S 3 3 bll,0 1957-1959 prices. Finally, as mentioned considerable growth has oc( urreci in (a,)jtui inputs from plant and equipment that are emplosecl iii Producing rjiI output but not owned by the railroad companies and therefore from the preceding capital inputs. During recent O\cI1I(Id ears, Outl,c On )rft freight cars by other than railroad Operating Companic's have be $300 to $500 million per year, or roughly running one-quarter to one-third the tot gross capital expenditures of the operating cOmpanies thenisefveç A lack of historical data on such outlays prevents arriving at an Cstimate of the total stock by cumulation so instead that value must be esti'nd b capitalizing the annual rental charges, as is clone to a rough approxin7j in Table 8. The ICC reports annual rental charges for equipment and iointl' Operfacilities, Of equipment rental charges, icc figures suggest roighlv 30 percent that represent maintenance leak ing th balanc to interest and deprecitjon As the hulk of these pay tgr equipment a 4 percent rate ot annual depreciationìrental charges are for ated 2S-ear lit) can assumed The average yield on equipment obligations about 3 Percent in 1947 and 6 percent in 1970, is used as the approprj,ite rate of interest These estimates imply that the capital stock that \ielded S 3 1 million in TABLE 8 Estimate of Total Value of Capital Stock Emplo'ed in Producing Rail Output but Ownrj External 1%' (millions) Iteni Hire o 2 S 4 5. Hire e. &\7uIpri1t,i. net tO tint rocil,t rents Iot, rent,)( ch.r' for 3 Annj,1j dt'pr Ancrj int'R S ° fl)lrntt'n,p,. - 034 5 t aho e 22 04 42 t'\tt'rl),7l\ o 'it'd rai1 lItliri r,it&' rate Cip4jI ret-ii 'r fOttor to apitol ,iIik' ifl rtntt fl turre dojLir R; 7 roj c i'r)tr tie lion ot 7 5 tor fllJlntenmce 01 p1ant 19-r equnitnt -I, pLint 3 i t ipOj7 \ alu' or r.'ntt' pL,it In 00 doll11 222 2 tfl lie 25 4 1,() to in tal all ed ng tal ck he by on r- at The U.S. Railroad Industry in the Post-World War II Period: A Profile 477 rental charges (net of maintenance) in 1947 had a value of approximately $1.9 billion, whereas the capital stock that yielded $557 million in rental charges in 1970 had a value of approximately $5.6 billion.° Converting these estimates to constant 1957-1959 dollars by the ICC's railroad construction cost index, the constant-dollar capital stock of externally owned rail plant and equipment can be estimated as growing from $2.83 billioii 1957-1959 dollars in 1947 to $4.56 billion in 1970. When inputs of both railroad-owned and externally owned plant and equipment are taken together, the striking aspect is the overall stability of the total capital stock used in the U.S. rail industry in the postwar years. Moreover, whatever increase has occurred, apparently has consisted largely of externally owned assets. In Table 9, the rate of growth of rail inputs and outputs, as calculated from the statistics just presented, are summarized. Crude estimates of the rates of growth of rail labor productivity and rail capital productivity can or or be determined by subtracting the rate of input growth from the rate of output growth. Labor productivity in the rail industry, by this measure )e in t. (using the figures of Table 9), grew at an average rate of 3.7 percent per year. Although this figure is substantially lower than the 5 to 6 percent rate of growth in labor productivity estimated by the BLS and Kendrick in their studies, it still is above the reported rate of growth in labor productivity in the entire private domestic economy, which has averaged 3.0 to 3.1 percent per year during these same years. This is a surprising record of accomplishment for an industry that has suffered declining output, with the result that increases in labor productivity have necessitated a rather rapid in contraction of the labor Iorce.' An important factor in this comparatively rapid growth of rail labor productivity is almost certainly the substitution that has occurred of capital TABLE 9 Rate of Growth of Rail Inputs and Outputs Percent Annual Rate of Growth Index Item 1947 1970 1947-1970 Output 100 84.0 Inputs Labor Materials and supplies 0.7 100 ioo ioo 36.2 45.7 100.2 172.4 4.4 3.4 ioo 104.3 0.2 Capitalrail-owned Capitalexternal Capilalrail-owned plus external 100 0.0 2.4 TABLE 10 Weightings for Aggregating Labor, Capital? and Intermediate inputs Rased on Contribution of Each to Total Rail Operating Costs, 1947 and 1 970 (all figures in millions of eurreizt dollars) 1947 I tern Labor (including payroll taxes & weltare benefits) Materials & supplies External capital Ra (-owned capital (excluding land) see Table 8 Depreciation Interest Land lnterest Total operating Costsa 1970 %of %of Amount Total Costs Amount Total Costs $4986 1,909 55 $6400 44 2 1,636 777 II 10 (65300) 1633 21 171 (34000) 850 1,020 (2,000) 60 $8,996 11 1 100 3,918 (3,500) 210 $14,574 5 11 27 1 100 tn the case of labor, purchased materials and is equated to actual supplies, and external capital, the contribution outlays. The cost of to total Costs to 2.5 percent of rail-owned capital inputs, however, is taken as depreciation reproduction cost less accrued depreciation to 3 percent in 1947 equal and 6 percent in 1970 of reproduction(in current dollars) plus cost less depreciatior', interest charge equal (in current dollars), for labor, as depicted strikingly in Table 10. Although output total capital inputs have has declined, actually terms a negative rate of capital increased slightly, implying in simplistic Rather than productivity growth.2 calculating either labor productivity the change in productivity for all factors could be or capital productivity, calculate the growth of multiple measured. In order to factor productivity, it is first determine with what weights necessary to combined. It is clear that labor, capital, and any other inputs should be the more heavily labor inputs (which declined) are weighted have relative to capital inputs (which have been increased slightly), the greater stable or will be the measured rate gain, and vice versa, The of productivity theoretically most defensible dure is probably weighting proceto weight each class of input to total costs; that is, the larger, more costly according to its contribution more heavily it should be a certain class of input, the weighted. The problem with as the proportion in which the various this strategy is that over the years, the proportion in which inputs are combined hs changed each input has presumably contributes to total costs changed also. Thus, the of observation. actual weights depend on the Fortunately, input year proportions measured in usually change in the opposite direction to their relative physical units prices (due to I The U.S. Railroad Industry in the Post-World War II Period: A Prolile 479 factor substitution), with the result that their proportions in total cost change less sharply than their proportions in physical units. The contributions of labor, capital, and intermediate inputs to total operating costs in the years 1947 and 1970 are shown in Table 11. sts Whereas the substitution among capital and intermediate inputs as against labor has moved steadily in one direction between those years, it may be assumed that the weightings developed using those two extremes are the extremes of the various weightings that would result from using intermediate years. Using the weighting scheme based on the input costs analysis for 1947, in which year labor and intermediate inputs were used rather intensively, multiple factor productivity in the rail industry grew at an average rate of 2.4 percent per year from 1947 to 1970; using the 1970 weights, in which year capital inputs were used rather heavily, overall rail productivity grew at an average rate of only 1 .5 percent per year. These estimates compare with a 2.5 percent rate for combined labor and capital productivity in the entire private domestic economy during this same period.43 In sum, the railroad industry may not be achieving the high rates of ual r). productivity growth often claimed for it. Much of the statistical gain in rail labor productivity owes simply to the decline of passenger traffic, of less-than-carload traffic, or traffic in manufactures, and of short-haul traffic, with the resulting shift in the composition of rail traffic toward heavyloading, bulk commodities, and longer hauls, which are inherently less d, tic ty, to to be ye or ity labor-intensive. Compensating for the changing character of traffic reduces the growth rate of labor productivity to a level that is close to the average throughout industry. This growth of labor productivity has also been accompanied by some replacement of labor inputs by capital and other inputs. Annual productivity growth for all factors in the rail industry has perhaps averaged only 1 .5 percent to 2.5 percent during the postwar period. This record is approximately the same or somewhat lower than the comparable productivity gain in most other industries, though not adverse compared to what might be expected in a declining inrtry c railroading.44 Ce- ion he at ed sts ear its to [4] THE RAILROAD FINANCIAL SITUATION REVISITED For a number of reasons, the standard reports on net railway operating income often are regarded by financial and other observers of the industry as being potentially misleading. They are, of course, prepared to ICC specifications and in many respects conform to accounting procedures found elsewhere in industry. However, some ICC procedures predate the 1972 1971 1970 1957 1968 1969 1965 1966 1962 1963 1964 961 1960 1935 1956 1957 1938 1959 1952 1953 1954 1951 1950 Year TABLE ii 486 696 835 962 1046 676 678 653 584 328 726 806 818 1128 1068 922 762 748 1040 943 1078 1109 874 Net Railway Operating Income 81 2. ROOe 790 707 732 756 773 766 655 699 686 641 629 553 582 602 614 539e 430 445 486 505 527 Depreciation Accruals I (,47 496 1276 1421 1432 1451 1669 1778 1213 1179 1381 1475 1304 1304 1364 1362 1621 1667 1401 1470 1388 1564 1614 Flow Cash Capital intact 178 1244 1351 1187 1509 1631 1953 1522 919 646 833 1044 1417 1395 738 818 1228 910 820 1260 1341 1066 1414 Capital Expenditures Gross 318 403 90 264 88 75 175 38 87 548 431 294 533 626 544 109 737 393 581 354 223 --26 404 Adjusted Net Railwa Operating Income (Cash Flow - Gross Capital Expenditures) I (,4 1 29 del. 1.07 deL def 0.15 0.35 1.75 2.23 2.17 1.20 2.21 2.55 0.44 1.60 3.08 2,36 1.44 0.91 def. 1.64 Returna Adjusted Rate of Railway Earnings Computed with Cash Flows Charged with Gross Capital Depreciation as the Cost of Maintaining Expenditures Rather than with 1 The U.S. Railroad Industry ri the Post-World War II Period: A Profile 481 emergence of modern accounting practices and are therefore at some variance with current accounting conventions. Unquestionably, the largest number of questions about railroad account- ing pertain to the handling of depreciation expenses or the capital asset accounts. For example, many suspect that the depreciation accruals, computed by standard ICC methods, underestimate the true degree of capital consumption in the industry. Other reservations about railroad depreciation figures originate in the fact that railroad equipment is assumed to have an uncommonly long average physical life by the standards of modern industry (usually in the range of twenty to sixty years). When coupled with the obvious fact of considerable inflation in costs in the U.S. economy over the last few decades, these long lives for depreciation could under certain assumptions (and particularly in a slow-growing industry)4s result in cash flows from depreciation that are inadequate to finance current replacement needs. The rail industry, as explained in section 1, has for some time experienced a stable volume of traffic and seems to have limited prospects for future growth. In such an environnient, it is difficult to believe that current gross capital expenditures would consistently run above a level needed to do more than maintain current output. In fact, the capital committed to the industry, as developed in the calculations reported in section 2, does indeed seem to have been more or less constant over the last couple of decades. More precisely, if one concentrates only on that capital owned by the industry itself and therefore subject to depreciation on the books of the industry, the total committed capital has actually declined somewhat in recent years. This suggests that a somewhat truer picture of the earnings situation in ON CcO the rail industry might he achieved by substituting gross capital expenditures for depreciation accruals in computing net railway operating income. The standard objections to such a procedure, applicable in almost any other industry, that gross capital expenditures not only Provide for the replacement of depreciating assets but also for additional assets necessary for the expansion of the business (and therefore cannot realistically be charged off as costs of business in the year incurred), would not necessarily be as applicable in an industry as stable (or even declining) as the railroad industry. Moreover, though gross capital expenditures in the U.S. railroad industry have run substantially above depreciation charges, there is some evidence that the quality of the equipment and property committed to railroading may actually have deteriorated somewhat in recent years; for example, there is much comment on a continuing so-called shortage of freight cars and an undeniable increase in the instances of derailnlent.46 Consequently, gross capital expenditures, as much in excess of depreciation charges as they are, may themselves be a conservative estimate of the S The U.S. Railroad Industry in the Post-World War II Period: A Profile a )f IC n St 483 investment in all but three years since 1964, even before fixed charges are pc',J. The financI viability of the industry depends ultimately, of course, on its ability to meei these fixed charges. The fixed charges of the railroad industry have increased markedly over the past decade, from $367 million as recently as 1963 to $601 million in 1971.48 The upward march of fixed charges owes to a combination of advancing interest rates in the capital market plus the increased debt of the railroad industry. Despite the stable volume of rail traffic, gross capital expenditures have exceeded internal cash flows (depreciation plus net income) less cash dividends in every year since 1960, thus requiring an addition to borrowed capital and greater al fixed charges. f Net income, the residuum of total income left after paying fixed charges 3 and miscellaneous deductions from income, is the amount of income e remaining for distribution to equity holders as dividends or for re- rn investment in the business. The gradual decline of total income and the rising level of fixed charges imply that net income has trended downward, as Table 13 indeed shows. Although net income has been falling, the total cash dividends paid out by Class I railroads have been increasing during the last decade, to the extent that dividends paid during 1970 and 1971 exceeded net income for the system as a whole. However, if the deficits in net income of the six northeastern railroads are added back into the net income of all Class I railroads, the dividend payout ratios for 1970 and 1971 are reduced to 68.7 percent and 57.8 percent for those years. Clearly, the financial condition of the railroad system as an entity is less than robust. The cash flows of the industry (net railway operating income plus depreciation) are barely sufficient to provide for capital maintenance if that is measured by gross capital expenditure rather than depreciation, let alone to provide a competitive rate of return on invested capital. Furthermore, cash flows have been stable or trending slightly downward, whereas the amount of money invested, and therefore the costs of amortization and interest, have been rising. An intriguing question is why monetary investment has continued at such a high or even expanding level in a declining or stagnant industry, earning such low rates of return. A high rate of reinvestment in an industry with a low average rate of return is understandable, of course, if the new investment is expected to earn a high marginal rate of return. Unfortunate- ly, continuing increases in the dollar value of railroad assets have been accompanied thus far by a fairly stable level of cash flows (railway operating income plus depreciation) and by a continuing low level or even decline in the average rate of return. The claim of high marginal returns can only be supported, then, by maintaining that although new investments earn a high return, the returns to old assets are falling so fast as to , 1956-1 06() 1966-1970 I 961-1965 Averages: 1960 1959 1958 1957 1956 1961 1965 1964 1963 1962 1968 1967 1966 19,9 1970 Year TABLE 13 13 16M 708.1 769,1 11473,2 762.3 922,2 1.068,2 584.0 961.5 818.2 805.7 725.7 537,8 485.9 654.7 677.6 676.4 1,045,9 (1) Net Railway Operating tncome 1 52.o 165.: 160,0 2,3908 147.1 50,8 157.6 155.6 153,8 159.5 160,4 161,1 159,7 159.6 170.3 165.4 165,1 163.4 161.4 (2) Road and Structure 603,6 511,3 442,7 7,788.7 470.9 458.3 448,5 430.7 405.4 41,2 525.2 507.8 495,6 628.8 609.5 607.6 601.2 570.9 (3) Equipment -Depreciation Sources__Raj I road Operations Funds Generated by InternaI 750,5 507.0 72,4 8149,8 439,j 627.1 286,7 196.5 312.6 749.8 448.8 313.1 224.2 799j 630.4 554.8 755.2 1,136,3 676.1 (4) Equipment Debt 142 77 273.2 2,765,7 159,9 214.8 45.6 137,3 753.1 151,4 184.5 166.8 59.8 126.3 50,4 333.0 512.6 369.8 100,4 (5) Funded Debt Sources 10.8 60.2 34.6 528.6 13.3 11.3 12.2 8.1 8.8 30.4 18.2 75.9 36.9 107.4 40,1 66.7 (6) Sales of Equi Funds Generated by External Flows of CapiI Fundc__(jass I Railroad Companies, 1956-i 970 (thousands) Funds Generated 423.6 478 0 707,4 8,060,4 179,0 513.5 434.2 732.8 614,0 307.6 285.2 430.8 4907 809.8 553.2 850.8 736.0 (7) Other Incomea Sources, lnctudjn by UnidentifiabI( 342.5 4730 5,636.4 289.3 275.6 346.3 314.6 325.6 322.3 365.4 368.9 482.4 505.3 520.6 457,5 (8) Other Income (continued) TABLE 13 455.1 415.8 419.7 424.5 410.6 402.6 6,675.1 1960 1959 1958 1957 1956 Totals: 1961 401.4 398.7 406.0 433.1 412.3 6,046.3 344.7 346.5 380.9 395.6 418.8 440.2 419.3 346,6 333.8 322.6 467,1 487.4 1965 1964 1963 1962 422.4 438.8 470.4 498.6 557.0 1970 1969 1968 1967 1966 545.3 505.5 (10) (9) Dividends Year teed Rents and Guaran- Interest i1 6.8 165. 160.0 152.9 603.6 511.3 442.7 3,990.6 331.9 303.8 135.5 144.9 273.8 258.7 326.7 222.2 108.7 262.6 340.9 324,8 570.7 155.0 230,4 (11) 6,254.5 391.7 368.0 379.3 355.3 342.9 422.8 516.4 402.0 373.7 352.7 541.2 530.3 335.9 492.8 449.5 (12) Total 21,653.3 1,212.5 1,361.6 1,364.6 1,503.7 1,620.7 1,503.8 1,474.6 1,381.0 1,178.6 1,668.1 1,285.0 1,429.6 1,450.3 1,441.0 1,773.2 (13) Fundsh -11,522.8 -1,079.0 -805.5 -622.5 -827.0 -1,026.1 -728.3 -378.6 -528.1 -700.5 -738.4 -833.2 -963.0 --850.7 -785,7 -656.2 (14) Fundsc Total Externally Generated 10.8 142.1 372.4 nternaily Generated 60.2 34.6 273.2 137.7 750.5 507.0 Equipment Funded Debt Debt Repayment Repayment Funds Paid to External Sources 708.1 769.1 1 966-1970 1961 -1965 1 956-I q6() 473.0 342.5 371.8 4,829.3 285.7 250.5 258,4 387.0 406.5 258.9 239.6 219.3 327.1 277,6 374.1 398.6 357.0 420.7 368.3 Road and Structure (15) 13,361.6 633.5 567.5 479.7 1,007.7 821.4 1,303.6 1,139.7 784.9 593.4 427.1 1,088.7 818.7 1,148.4 1,554.2 993.1 (16) Equipment -Capital Expendituresd_ 707.4 425.6 478.9 and Guaran- Interest 459.5 372.5 377, Interstate Commerce Commission 410,3 414.6 510.1 (10) Dividends 324.4 235.8 238.0 (11) 469.9 413.5 367.4 (12) 1,476.8 1,441,2 1,412,6 (13) Generated Fundsb Internally Total 679.9 752.6 872.0 Furidse (14) Total Externally Generated 317.6 2645 383.7 (15) Road and Structure 1,120,6 849.7 702.0 (16) Equipment Capital Expenditures_. Transport Stat,s:,cs in the United Ste, Table reproduced from pane No. 271, pp. 54 and 55, eriiied statement of Dr R A Nelson before the CC in hear,ns on Es Column 7 comprlces sources of funds that cannot be identified ri ICC summary data. It probably Income from non.rajl sources, includes proceeds from sale of railroad propeny, shon-term borrowing, and 5The sum of Columns 1, 2, and 3 equals column 13 The sum of columns 4, 5, and 6 less the sum of columns 9, 10, 11, and 12 equals column 14 drhe sum of columns 7, 3, and 14 equals the sum of columns 15 and 16 SOURCE. 1966-1970 1961-1965 1956-1960 Averages: teed Rents Equipment Funded Debt Debt Repayment Repayment Funds Paid to External Sources (concluded) Year(9) TABLE 13 I The U.S. Railroad Industry in the Post-World War II Period: A Profile 487 offset the impact of the new investments on the average rate of return.50 But beyond some point, the erosion of the return on aging assets can no longer be ignored in claiming a high marginal return on new investments. It is therefore significant that even if all the earnings of all the capital employed in the rail industry (i.e., net railway operating income) were attributed to the gross capital expenditures of the preceding ten years alone, the rate of return in 1970 would have been only 3.7 percent, in 1971, 5.2 percent, and in 1972, 5.9 percent. These rates, although higher than overall rates for the rail industry, still are low compared to industry in general. Continuing reinvestment in railroad property also may reflect the absence of alternative uses for the funds available to the railroads. Regulation, in particular, has impeded withdrawal of funds or diversification "out" of railroading. The common-carrier obligation of railroads enforced by the ICC and the specific prohibition of line abandonment and related forms of disinvestment may sometimes have caused the railroads to reinvest despite low anticipated returns.51 Given that railroad management is committed (whether by choice or by regulatory compulsion) to remaining in the traditional railroad business, simply doing so necessitates a continuing high level of investment.52 If rail service is to be provided, new cars must be purchased as old cars wear out and as traffic grows, and so on. The rising monetary cost of replacements in an inflationary economy contributes to the growing dollar investment in railroad property. Furthermore, railroad equipment is depreciated on its original cost rather than its cost of reproduction, arid the service lives of railroad assets are comparatively long, so the cost of replacing them is usually considerably in excess of depreciation based on historical costs. This may be offset to some extent by technological innovations, but it is doubtful if this offset is total or complete for the railroad industry. Of course, if the sum of gross capital expenditures plus cash dividends exceeds cash flows (net income plus depreciation), an industry must rely on external funds. Table 13 shows the volume of external funds that the rail industry has succeeded in attracting in recent years.63 In view of the chronically low average rate of return on capital invested in railroad property and of the apparent withdrawal of equity capital through dividends, this ability to raise outside debt is rather remarkable. One apparent explanation for the relative ease with which external funds have been raised is to be found in their form. The largest category of external funds shown in Table 13 is "equipment debt." Equipment debt comes in two primary categories, equipment trusts and conditional sale contracts (or deferred payment contracts).54 In cases in which neither equipment trusts nor conditional sale contracts are feasible, the long-term lease provides a third method of raising external funds for rolling stock. Equipment debt has been made available to the railroad industry be- 488 John R. Meyer and Alexander I cause of the near perfect security or collateral that the equipment to the creditor. "A history of equipment trusts," in the words of Oneprovides Cxper "indicates practically zero risk to holders. The Security is readily repo5 sessed and, if necessary, resold. Considering that the car life is more Iikef 30 years than 15, the market value from date of issue is greater than the indebtedness generally. Thus there is no risk premium. Another attractive feature of equipment debt is the seniority of the claims on earnings of the creditors compared to the claims of bondholders and stockholders The power of repossession puts the claims of those underwriting equiprn00, trusts and conditional sale contracts ahead of those holding the funded debi Together, the seniority of claims and the liquidity and mobility of the collateral have provided the railroads with a large source of external funds in equipment debt; these funds, moreover, have been made avail. able to the railroad industry at very favorable interest rates Interest rates on rail equipment obligations have tended to be equal or slightly lower than yields on AAA corporate bonds and rates on conditional sale agreements have been only slightly less than half a point highcr.' One not totally incidental consequence of the substitution of equipme debt for funded debt and equity capital is that it has already substantially complicated the problem of reorganizing railroads in bankruptcy Capital costs as a proportion of total railway operating "expenses" in the years 1925, 1935, and 1970 are shown in Figure 1. Traditionally reorganiza. tions have succeeded by changing the capital Structure to reduce debt service costs to a point at which operating revenues can reasonably be expected to cover them. Equipment debt cannot necessarily he written down in this way, if the creditors prefer to exercise their option to repossess. Thus it is only the claims of holders of funded debt and of lessors of fixed plant that can be reduced in this manner. Figure 1 shows that, as a result of the substitution of equipment debt for funded debt and equity, the proportion of capital costs that might be easily voided by bankruptcy has declined greatly. Compared to 11.4 percent of total operating costs in 1925 and 18.5 percent in 1935, readily voidable capital costs for the industry in 1970 comprised only 3.0 percent of total operating expenses Hence, the bankruptcy courts presently have a much narrower margin within which to work than during earlier rounds of railroad bankruptcies. In short, the railroad industry has continued to maintain its investment in railroad Property and has managed to attract new external debt to fund these capital expenditures, despite a low rate of return on invested capital. In essence, the industry has to Some extent substituted debt for equity and leases and conditional sales agreemen for debt in its financial structure. in the long run, obviously, some limit must exist to the extent to which an industry can attract relatively cheap capital by using more senior forms of financing thereby subordinating older debt. FIGURE 1 Capital "Rents" (Excluding Equity Capital) Percentage of total cost (revenue less profit and loses) 20 18 16 14 12 Fixed plant rentals 10 Interest--other debt 8 6 Net equipment rentals 4 2 Equipment obligations 1925 SOURCE; 1935 1970 Association of American Railroads [5J SUMMARY The evolution of the freight market has created serious problems for the railroad industry. Aggregate intercity freight traffic (of alt carriers) has grown less than two-thirds as fast as real GNP during the postwar period, and traffic of a type suited to conventional rail transport has grown still more slowly. The emergence of new carriers, particularly trucking and 4tJ john R, Meyer and Aiexander L. Morton pipeline transport, have narrowed the number of freight markets in rail service holds a cost and service whith advantage, further circumscribing growth. Conventional and widely used measures of railroad productivity, as "ton-miles per man-hour," such indicate that rail productivity has rate of 5-6 percent a year during recent grown at a decades, considerably average 3.0 percent growth of labor productivity above the in the private during these same decades. However, econony using alternative measures (e.g., allowing for changes assumptions and in the composition of rail can be argued that the growth in traffic), it rail-labor productivity has been about 3.7 percent. Capital inputs only to the railroad industry have nearly so rapidly as labor inputs, not declined so that the indicated crude growth rate for rail capital is near zero. When labor, capital,productivity inputs are weighted and other together, productivity in the industry only about 1.5 to 2.5 percent may have grown ity growth rate at or slightly per year during recent decades, a productivbelow the 2.5 percent per private economy as a whole. year rate for the The financial condition of the growth of traffic and productivity. railroad industry reflects the sluggish On many grounds, or so overall average rate of return on transportation in fact, the 3 percent U.S. Class I railroads, using operations reported by conventional ICC accounting be an techniques, may overstatement. Trends in railroad operating income and capital expenditures also point toward a possible further railroads' financial condition. deterioration in the In short, the U.S. railroad industry suffers severe problems. necessarily mean, though, that the industry can never be This does not efficient, restored as an competitive, profitable industry.60 Indeed, important segments of the even today, there are reason to expect continuedindustry that are progressing and have every reasonable level of prosperity prosperity. Nevertheless, attainment of a for the industry require quite as a whole probably will dramatic changes, either in structure of the industry, current economic trends or the or both. Total cash dividends Equipment and facility rents Net railway operating income Other income Total income Total fixed charges Income after fixed charges Ordinary incomea Cash flow Net income Net railway operating income before federal income taxes Ordinary income before federal income taxes Netoperatingrevenue Railwaytaxaccruals Railway operating income Total operating revenue Freightrevenue Passengerrevenue Mail revenue Total operating expenses Maintenance: way and structure Maintenance: equipment Transportation Operating ratio (percent) Item 90.1 780,227 100.0 100.0 100.0 786,919 647,543 385,493 123.2 126.0 98.2 969,768 815,746 378,549 112.1 112.4 95.5 882,483 727,803 368,164 96.5 92.8 624,900 357,561 612705 99.2 92.4 99.0 86.7 86.0 92.1 93.1 77.95 571,017 1,023,229 100.5 101.5 91.9 102.2 98.5 99.3 98.4 98.4 98.0 108.2 88.7 128.5 113.6 138.0 95.3 122.1 98.7 140.3 146.6 102.9 Index 2,107,874 686,387 1,221,487 415,828 805,659 330,074 1,135,733 367,970 694,065 651,637 1,226,326 99,1 96.9 98.1 103.7 $9,559,522 8,146,131 588,104 338,684 7,451,648 1,182,507 1,731,735 3,771,254 1963 Amount 98.0 98.8 103.7 90.6 117.5 106.7 124.3 94.0 113.0 98.4 123.8 128.4 85.9 96.2 93.8 95.7 96.8 99.5 98.3 99.2 97.2 105.4 103.1 537,771 322,281 860,052 369,111 428,976 382,444 1,073,155 99.2 99.6 $9,439,895 7,991,146 619,056 343,588 7,418,562 1,154,802 1,743,639 3,755,092 78.59 2,021,333 905,044 116,289 390,610 725,679 325,575 1,051,254 366,817 96.6 96.6 97.6 $9,189,138 7,739,044 624,688 341,697 7,274,260 1,117,680 1,683,363 3,710,832 79.16 1,914,878 991,083 923,794 386,023 95.1 Index Amount Index 1962 Amount 79.52 1,948,958 998,799 950,159 366,143 584,016 346,328 930,344 372,751 494,799 444,640 1,191,693 Index 1961 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 $9,514,294 8.025,423 640,268 331,378 7,565,336 1,191,690 1,759,828 3,832,882 1960 Amount APPENDIX TABLE A Income Statement, Class I Railroads, 1960-1 970 (dollar amounts in thousands; index, 1960 = 100) (cont;nued) APPENDIX TABLE A (conc! ed) Total operating revenue Freight revenue Passenger revenue Mail revenue Total operating expenses Maintenance, way and structure Maintenance; equipment Transportation Operating ratio (percent) Net operating revenue Railway tax dccruals Railway operating income Equipment and facility rents Net railway operating income Other income Total income Total fixed charges Income after fixed charges Ordinary incomea Cash flow Net income Net railway operating income before federal income taxes Ordinary income before federal income takes Total ca.h dividend,, Item APPENDIX TABLE A 121.5 129.1 956,131 836,103 457,188 118.6 157.0 102.1 149.4 27.6 106.5 140.1 108.7 87.2 131.4 117.4 98,7 116.1 103,6 105,4 90.3 99,4 102.3 102.8 100.2 102.3 Index $9,856,527 8,455,457 577,910 329,169 ,737,847 1,225,759 1,763,786 3,920,622 78.50 2,118,679 870,581 1,248,098 429,836 818,213 368,891 1,187103 380,420 739,106 698,184 1,383,824 Amount 976,285 470,800 1,125,172 1,326 905 400,565 855,607 814,629 1,521,210 1,441515 479,ggg 961,516 365,389 311,341 7,849,841 1,235,801 1,774,878 4,020,161 76.90 2,358,009 916,494 $10,207 850 8,835,958 553,056 Amount 22,1 5i, I43Q 103.8 103.7 100.8 104.9 96.7 121.0 91.8 151.7 131,1 164.6 105.5 142.6 107.5 172.9 183.2 127.6 86.4 94.0 107.3 110.1 Index 1,090,108 499.364 1,232,785 1,303,739 1,843,089 4,139,268 76.19 2,537,010 968,372 1,568,638 522,775 1,045,863 399,492 1,445,355 425,304 941,489 903,783 1,634,349 8,117657 $10,654,555 9,280,513 543,632 303,999 Amount I 28. '1 168,3 156.6 114,2 190.3 203.3 137.1 11 5,4 1 55,4 112,0 115.6 84.9 91.7 107,3 109.4 104,7 108,0 95.8 130,2 97.0 165,1 142.8 179.1 Index 196 6-_ 620,7'j 742,752 1,310,1)92 321,541 676,434 457,543 1,133.980 460,923 568,774 553,789 910 178 1,251 371 574,936 1 11)4 95.8 94, 121.9 123.6 119.0 124.5 109.9 132.1 115,8 1 57.0 91.1 131.7 99.6 110.9 2,161549 109.0 113.8 75.8 69.6 103.4 108. I 106,1 109,2 Index 485,359 263,533 8,204,492 1,287,834 1,867,789 4,186,049 79,i 9,1 30,233 $10,366,041 Amount -1 967___ I SOURCE: ICC, AAR, Moody's transportation manual. Prior to 1967 reported as net income. Total cash dividends Cash flow Net income Net railway operating income before federal income taxes Ordinary income before federal income taxes !ncomeafterfixedcharges Ordinary income' Mail revenue Total operating expenses Maintenance: way and structure Maintenance: equipment Transportation Operating ratio (percent) Net operating revenue Railway tax accruals Railway operating income Equipment and facility rents Net railway operating income Other income Total income Total fixed charges Passenger revenue Total operating revenue Freight revenue I tern APPENDIX TABLE A (concluded) Total cash dividends 118.6 1968- 133.7 98.1 635,480 515,590 94,5 114.4 128,1 69.4 59.0 113.4 117.9 108.8 113.6 99.4 116.7 94.7 139.7 177.5 116.0 144.3 126.5 129.8 121.8 114.1 121.5 Index 1969 Amount 122.1 487,440 620,429 760,861 1,502,958 2,002,316 4,595,574 79.18 2,383,796 1,029,067 1,354,729 700,059 654,670 505,267 1,159,937 521,346 545,054 514,238 1,280,043 463,565 9066,529 77,284 10,346258 438,667 $1 1.450,325 470,800 743,702 2,273,718 946,334 1,327.384 649,760 677,624 499,639 1,177,263 483.815 602,779 569,402 1,363,132 564,505 79.05 1,914,265 4,354,705 1405,132 195,418 8,580,961 $10,854,678 9,749,788 444,334 Amount 457,188 126.4 95.8 96.7 120.3 128.9 68.5 53.5 119.8 126.1 113.8 119.9 99.6 122.3 103.0 142.5 191.2 112.1 145.9 124.6 139.9 110.1 115.7 107.4 index 499,364 1970 Amount 502,872 421226 314,933 574,204 1,068,518 1,263,158 777,304 485,854 482,433 968,787 589,148 256,851 226,583 1,016,177 76,179 2,331676 9659,9:32 1,612,535 2,165,234 4,873,299 80.56 420452 161437 $11,991,658 10,921,813 129.5 109.3 48.7 73.0 101.3 119.6 107.0 132.9 212.3 83.2 139.3 104.1 158.1 51.9 51.0 85.3 127.1 65.7 48.7 127.7 135.3 123.0 136.1 126.0 Index 130.4 p (j 494 John R. Meyer rind Alexander L. Mort0n NOTES The income statements a Class I railroads combined fur tlit' l% to 1970 are given in Appendix Table A. Operating income is computed net of payments for hired equipment and joint fasjlisj05 Payments for hired equipment have risen rapidly during the past decade from $3 million in 1960 to $816 million in 1971 cf. section 4>. Net investment represents original cost, less depreciation and amortization, accrued under the accounting regulations of the ICC. The return on investment is consistent with respect to hired equipment (e.g., private cars> in that payments for hired equipment are charged off as expenses before net operating income is computed, while the value I this equipment is not included in the net investment accounts. The effect of this partial exclusion of capital is to leverage the rate of return shown in Table more highl> The charge is frequently leveled that the net investment figures shown in Table I are inappropriate as a rate base because they include economically obsolete assets and much over-valued real estate. In defense of the net investment figure, it shoijl] be noted (1) that the most blatant cases of historical over-valuation h-ave been removed b write-downs in a series of bankruptcy proceedings that swept the railroad inductry in the thirties; (2) that in 1963 the commission (1CC> issued an order which required all railroads to adjust their corporate books and substitute the adequately supported cost of property figures shown in the valuation records in place of the rejected historical Costs; and (3) that gross capital expenditures on rpadway structures and equipment between 1947 and 1972 alone have totaled $30.8 billion, or more than 85 oercent of the gross investment in transportation property recorded on the general balance sheet for 1972 See 3391CC 164, in which the commission defends the existing valuation of assets 1 asa rate base. First Natsonal City Bank of New York, Monthly Letter, April 1973. Both net railway operating income and net investment in transportation property (and,a fortiori, rate of return on investment) are affected by the fact that most railroads use retirement accounting rather than depreciation accounting to writeoff certain categories of investments in road and structure, such as grading, tunnels, rails, ties, and ballast With retirement accounting, these assets are carried on the books at full value (original cost) and written off only when they are retired, Replacements of rail, ties, ballast, and other assets subject to retirement accounting are charged to operating expenses; ie., written off in the year in which the expenses are incurred. Without detailed analysis, it is not possible to say svhether the use of retirement accounting results in net railway operating income and net investment being higher or lower in the long run than they would be if conventional depreciation accounting were used. It is probable, whichever is the case, that the level of net operating income and the level of net investment would be changed in the same directron, so that their quot:ent (rate of return on investment) would not be much affected in the long run by the choice of accounting technique. One railroad, the Chicago and Northwestern, recently conaccounting: reported net income during the first eight months under depreciation accounting was about $5 million higher than it would have been verted from retirement accounting to depreciation under retirement accounting. That is. the increase in deprecia- tion accruals was more than offset by the decline in expenses as outlays on track ceased to be recorded as current expenses. However, there was a major write-down in asset value at the time of conversion from retirement accounting to depreciation accounting so that depreciation accruals increased less than they would have otherwise One effect of retirement accounting is to give management greater control over te level of profits reported in any particular year. linlike depreciation costs. which accrue automatically each year based on historical investments, the recorded "cost" of ssea'- The U.S. Railroad Industry in the Post-World War U Period: A Profile Ortoq 495 and-tear to assets subject to retirement accounting in any particular year is essential!y the replacement expenditires in that particular year. So management can control or vary the annual cost of assets subject to retirement accounting in the short run by varying the level of replacement expenses. A large number of assets do not require periodic replacement and are carried on the railroads books at original cost. One financial analyst estimates that the rail industry carries about $3.5 billion worth of gradings and tunnels on its books in this way and that railroad cash flows would increase by $65 to $75 million annually were Congress to permit the railroads to amortize these investments and, thereby, reduce their income tax 70 are aciljties m $321 'ICCrued eflt with ent are at Costs; liabilities. See Pierre S. Bretey. Railroad Industry Review (New York: Reynolds Securities, Inc., 1972), p.29. Of course, for railroads operating ala loss or having no tax liability for other reasons, amortization of presently nondepreciable assets would have no material value. See Alexander L. Morton, 'Freight Demand," Ph.D. dissertation, Harvard University, 1973 (unpublished), Chapter 5. 'Sand, gravel, and stone" is actually the largest category of raw material consumption in terms of weight, as Table I indicates, and has also been the fastest growing during the postwar period, owing to the acceleration of highway construction, which uses these materials intensiveiy. However, sand, gravel, and stone for direct use in construction tend to move very short distances in large volumes and to rely on specialized private carriage, thereby making disproportionately small demands on the general transport between system. value of S partial highly. e 1 are sets and be noted oved by ry in the uired all cost of The case of food illustrates this point. Although per capita expenditures on food rose from $405 in 1950 to $491 in 1970 (in constant 1967 dollars), the weight of food consumed per capita, as measured by the U.S. Department of Agriculture, fell from the gross or 1972. ssets as a 1,506 pounds in 1950 to 1,449 pounds in 1970. This increasing decentralization and market orientation of production is documented in studies by B. Chinitz and V. Fuchs. See 13. Chinitz, Freight and the Metropolis (Cambridge: Harvard University Press, 1960), pp. 114-115; and V. Fuchs, Changes in the Location of Manufacturing in the United States Since 1929 (New Haven: Yale University Press, 1962), pp. 292-293. In many instances some form of processing of the raw material occurs at or ty (and,a ads use ategories d ballast near the raw material source, but the resulting intermediate good to be transported is usually still in bulk commodity form. Freight rates on bulk commodities have declined more rapidly than (or not increased so much as) those on manufactures (see Chinitz, 1960 (Note 9), pp. 119-120). This owes in part to the lower labor intensity of transporting bulk commodities, and in part to the practice of value-of-service pricing, which encourages a greater mark-up over costs on (original last, and ises; i.e., reti rerrieni higher or the higher valued manufactures. ing were Morton, 1973 (see note 7), Table X, p. 91. Morton, 1973 (see note 7), Table X, p. 91. For example, in producing automobiles, machinery, instruments, and other metal products, metals may be transported as many as three or four times in one form or another (ore, sheet metal, stamping, subassembly) before leaving the final factory as a e and the quotient the choice ently conet inCOilie on higher deprecinack ceased o in asset ccoUfltiflg, vise. l over the ch arc we of wear- - finished product. Morton, 1973 (see Note 7), Chapter 3. The mechanization, computerization, and "routinization" of industrial production and distribution raise the allowable premium for reliable del:veries. In food distribution, for example, handling the routine shipment has become almost costless to the consignee compared to the lost, delayed, damaged, or otherwise exceptional shipment. See L. W. Weiss, "The Survival Technique and the Extent of Suboptimal Capacity,"loumal of Political Economy, June 1964, pp. 246-261, especially Table 1, p. 249. This tendency for regional economies to acquire their own production capacity in a 0 John R. Meyer an(IAk'xander[ 496 Mort00 wider variety of manufacture as the size of the market expands is known in international trade a "import substitution.' Not only are individual cities becoming larger, but urban growth seems to 1w' qtrated in a limited number of metropolitan belts, such as those extending from Bst10 to Norfolk, from San Francisco to San Diego, from Miami to Jacksonville and around the. Great Lakes. These four "nwgalopoli" alone have absorbed almost all the nation's net population growth during recent decades; in 197() they included 41.1 percent of th population and it is estimated they svill include 60.1 percent by 2000. (See Jerome. p Pickard, Dimensions of Metropolrtanism, Urban Land Institute Research Monograph 14 Washington, D.C., 1967.) As this type of agglomeration continues, even a limited nunnJ1 of plants producing any one item will be able to supply a majority of consumers Without resorting to long hauls to distant markets. Advisory Commission on Intergovernmental Relations, tjrhan and Rtira/ Arnrin-a. for Future Growth, Washington, D.C., April 1968, p. 3. Poiic5 See J. R. Meyer, I. F. Kain, and M. WohI, The Urban Transportation Problem (Cambridge Harvard University Press, 1965), Chapter 2; J. F. Kain, "The E)istributiori and Movenienioi Jobs and Industry,'' The Metropolitan Enigma: !fl(;uirieS into tlic' Nature ,ineJ Di;flension ( America's "Urban Crisis," James Q. Wilson, cd. (Cambridge: Harvard Universit Press, 1968); and Edwin Mills, Studies of the Structure of the Urban Economy (Baltimore Hopkins Press, 1972), p. 35. Johns Raw materials and semi-processed goods were brought into cities by rail to factories clustered around the railhead. The products of these factories were loaded directly back Onto railcars for distribution to consumers in other cities and rural markets Manufac. lured goods arriving by rail in cities arid towns were easily distributed to consumers who generally resided relatively close to the business core and railhi'ad. The rate of suburbanization is not uniform among manufacturing industric's. light manufacturing has suburbanized the most rapidly since it is least concerned with high-volume throughputs of commodities (other than fuel and electricity), hence. least dependent on rail service for delivery of inputs and least likely to employ rail service fct delivery of outputs. But it is also light industry that is growing fastest among manufacti,,r. ing industries and producing the greatest growth of freight traffic (in terms of revenue> Thus, suburbanization handicaps the railroads in participating in the faster-growing sectors of the freight market. The growth rate of rail labor productivity calculated in this fashion is, in fact, the eighth highest among the growth rates for 35 individual industries calculated by the BLS for the period 1957-1970, See hidexes of Output Per Man-Hour, Selected Industries, 1939 and 1947-1970, 31.5 Bulletin 1692, p. 7. 24, John W. Kendrick, Postwar Prodoitis'ity Trends in the United rates, 1948.-1969 (Ness York. National Bureau of Economic Research, 1971), Chapter 5. Output ncr man-hourrose 8.2 percent a year in air transportation and 9.1 percent a year in pipeline transport, however, during this same period, Kendrick, 1971 (see Note 24), p. 5-8. Non-railroad freight cars account for an even greater proportion of total investment in rail freight cars, as private ownership tends to be cunc entrat('d anlur,g higher-priced tank cars, rc'frigerator cars, and cars of special design. One other consideration that would reinforce this bias would be that tht' (JualitV of each man-hour has probably improved over the years because of rising educational levels, etc., though in the case of railroading this may have been li'ss than in inrlustry in general because of the relative aging of IFie rail labor force. Deferring one expense may cause another offsetting the reduction in total costs that r:ategn)ry of c'xpenst' to rise, thereby partially can be hieved by th'fc'rral. For example, 1 he U.S. Railroad Industry ii' the Post-'A'orld 'A/ar II Period: A Profile 497 tonal deferred replacement of equipment causes higher equipment maintenance expenses; deferral of track maintenance necessitates slow orders" that escalate train-crew hours, etc. the ICC first assigns expenses that are 'solely retated" to either freight or passenger service to their respective service; then common expenses are apportioned statistically. "Solely related" passenger service expenses include, therefore, only those expenses that I the could be directly avoided if passenger operations were ceased; it does not inilude e elements of track maintenance, signal system operations, etc., that svoutd be incurred anyway, though almost surety in lesser amounts if passenger service were discontinued. Assigning only solely related passenger expenses to passenger service and all remaining 14 oh) how operating expenses to freight service, the average cost of passenger Service per ici5 passenger-mile was 6.0 times as great as the average cost of freight service per ton-mile C in 1963, and 6.75 times as great in 1970. On a "fully-allocated" basis, on the other hand, passenger service was 8.5 times as expensive as freight service in 1963 and 8.8 times as costly in 1970. The index of railroad carload freight rates prepared by the ICC lRI-ll indicates that carload rates rose 47.5 percent betsveen 1947 and 1970. Yet average revenue per ton-mile for all freight rose only 32.7 percent during this same period. The difference between these two figures is a rough indication of the extent to which the coniposition of rail freight traffic is shifting toward movements whose rates are tower, presumably because they are less costly to produce. Although this presumption is clearly more ci e wSS, ries dck fac- ho - ight ith ast for r- e). ing hth the nd ew se rI. in nk h Is, ral fly 31. accurate for different classes of freight service than for passenger vs. freight weightings, it is still of limited validity because of value-of-service rate structures. Kendrick, 1971 (see Note 24), Chapter 5. John W. Kendrick, Productivity Trends in U.S. Transportation Industr;es, prepared for the Office of the Under-Secretary for Transportation, U.S. Department oUommerce, January 1966, p. 22. 33. In genera!, as the Kendrick calculations illustrate, many factors influence productivity and complicate its measurement. At best, productivity measures incorporate an element of art as well as of science, For an excellent summation of the "state of this art" see S. Fabricant, "Perspectives on Productivity Research," prepared for the Conference on an Agenda for Economic Research on Prodoctivity, Washington, D.C., April 6, 1973. Sponsored by the National Commission on Productivity, 34. Rail ton-mileage fluctuates with the business cycle, and these cyclical variations are large compared to the secular growth in ton-mileage. Consequently, the measured growth of rail output varies with the choice of terminal years. Cyclical fluctuations during the forties were sizable, The 658 billion ton-miles of 1948 represents something of a compromise between the wartime high of 741 billion ton-miles in 1944 and a postwar low of 529 billion ton-miles in the 1949 recession. Were 1949 used as an initial year rather than 1947, rail output (freight and passenger; would have grown hy 0.4 percent per year (1949-1970) rather than declining by 0.7 percent per year (19471970). But this is to measure freight growth from a rather deep reessionary losv to a near all-time peak, At 765 billion, ton-mileage in 1970 was off 3 billion from its former record of 768 billion in 1969, but wefl above the 740 billion ton-miles of the 1971 recession, If the new record of 778 billion ton-miles in 1972 were used, measured rail output (passenger and freight) would decline by 0.6 percent per year (1947-1972; rather than by 0.7 percent per year (1947-1970;. 35. Man-hours paid for have declined at a rate about one-tenth of a percentage point less than the annual rate at which man-hours worked has declined during the postwar period. Also, it should again be noted, as in Note 27 above, 32. that the quality of S 498 John R. Meyer and Alexancltr '. Mo man-hours may well have improved over time so that even ''n-on-Jm worked" so unambiguous a term as it might at first seem. See U.S. Department of Commerce, Survey of Current Business june 1956 and Marc)1 issue for 1958 and all subsequent years. This series closely parallels the gross capjta expenditures reported by Class I railroads through 967. Beginning with 1963 Ilk Department of Commerce figures appear to include capital exp(flcljturrs Ofl railr()a() equipment by other than railroad companies, so in these latter years the gross capital expenditures reported by Class I railroads have been used instead in order to keep the. series consistent. The series of gross capital expenditures may understate actual investment in rail plant and equipment. Because of the use of retirement accounting rather than deprecj05 accounting, replacement and renewal of rail, ties, ballast, etc., are charged to curreni expenses and so do not appear in gross capital expenditures, This Possibility will be invoked in the next section where it will be suggester) that capital expenditures may be a more accurate measure of the actual depreciaj0 grO5S than the figures shown in the railroads' depreciation accounts, which reflect original acqujs1. lion values. Jack Fawcelt Associations, Inc., CapitalStock Measures for Transportation preparedforthe Office of the Secretary, U.S. Department of Transportation, June 1972, Vol. I, p These estimates are at some variance with the value of freight cars owned by ship1 and private car lines, which make up most of the external capital. The net investment (purchase price less accrued depreciation) in private cars was only $3.2 billion in and $0.3 billion in 1947, according to ICC accounts. See ICC, Transport Statistics 1971 United States, Part 9, 1971; and ICC, Selected Statistics from in the the Annual Reports of Owner5 of Private Cars, 1947. Another recent study of employment and labor productivity in the rail industry proceeds by a different method of analysis to conclusions similar to those reported here. (Paul H. Banner, The Measurement of Productivity in Rail Transportation a paper for presenta. tion at the 1973 annual meeting of the American Society of Mechanical Engineers.) That study begins by questioning whether the ton-mile is the relevant measure of output with regard to railway labor. For example, does one wish to assert that because the movement of a freight car loaded with 100 tons of coal rather than 25 tons of computer equipment generates four times as many net ton-miles for each mile it travels, it represents four times the productivity? In terms of the work effort involved, carloads car-miles, and train-miles are the most appropriate units in which to measure ssork output in the rail industry. Man-hours of work are disaggregated by employment category, and each category of labor output is assigned to that unit of output that is most relevant to its task. Thus, for example, "freight station foremen, laborers on platforms. yardmasters are associated with carloads. . freight engineers are assigned to train-miles... - Road maintenance, train dispatchers, Claim agents are assumed to vary snith car-miles on the assumption that loss and damage is correlated with distance." (Banner (see above), pp. 8-9.) Professional, supervisory executive, and similar activities are assigned to "overhead labor," which is in turn assocjatj with a weighted compasrte index of car-miles, train-miles, and carloads The growth in labor Productivity is then calculate! as the growth in carloads, car-miles and train-miles per man-hour of associated labor. This procedure states, in effect, that result of increases in the average capacity of individualincreases in ton-miles that are the cars, the average number of cars per train, the average length of haul per car, etc., do not reflect an increase in the outrsjt of numerous categories of railsvay labor and should not be attributed to those categories in computing their productivity The average annual rate of growth in labor productistt during the period 1946-1971 in each of the four output diniensions is as follosss S I Orton The U.S. Railroad Industry in the Post-World War II Period: A Profile 499 is arch apital 8, the ilroad apital 'P the plant lation urrent Rail labor associated with carloads Rail labor associated with car-miles Rail labor asociatd siih train-miles Ovrliead labor associaied wits a composiie index 2.5ct 2.7% 4.1% 1.5% If the four divisions of the rail labor force are combined according to their respective shares in total labor compensation, the average annual rate of labor productivity growth for all railway labor during the twenty-live year period is 2.7 percent per year (Banner (see above>, p. 10). The low rate of gain in capital productivity may be seen by comparing output to the principal forms of capital inputs: freight cars, locomotives, road, and yards. Revenue freight ton-miles per ton of freight-car capacity increased from 6,280 in 1947 to 6,400 in stmenl 1970, an average gain of on'y 0.1 percent per year. Revenue tonnage originated per ton of freight-car capacity actually declined 16 percent from 14.7 tons in 1947 to 12.4 tons in 1970. Ton-miles per locomotive in service rose at an average rate of 2.6 percent per year during this period. Taking miles of road as a proxy for capital investment in railroad lines other than yards, ton-miles per mile of road grew at an average rate of 0.8 percent per year. Taking miles of track in yards and sidings as a proxy for investment in yard plant, freight ton-miles per mile of yard track and siding rose 1.0 percent per year. In short, it is apparent that capital inputs have not contracted so freely as labor inputs so as to produce productivity gains. 1971 The choice of terminal years for calculating productivity growth affects the results :n the somewhat, but not severely. For example, using the record-high year for freight, 1972, rather than 1970. the growth of total rail output is increased by 0.1 percent per year, and gross than quisifor the 3-49. uPPers wners so total productivity growth is also increased by 0.1 percent per year. The most >ceeds aul H. senta1 That pt with se the pputer els, it 'loads, work yment s most ornis, (chers, , with anner es are posite then tur of ie the cars output gories ctivity lows: favorable comparison possible is to measure growth from the rather deep recessionary iow of 1949 to the all-time high of 1972. Over this interval, measured rail output growth is 1.0 percent higher, so measured productivity would be almost 1.0 percent higher (i.e., reduced somewhat by a rather sharp year-to-year drop in employment in 1949). A recent study of differential productivity gains among industries concludes that there is a significant positive correlation between productivity gains and output growth. Cf. Kendrick, 1971 (see Note 24), Chapter 6. "The relationship is reciprocal: relative advances in output affect productivity through differential scale economies; and relative changes in productivity, mirrored in relative changes in prices of the outputs of the various industries, in turn affect relative changes in sales and output." For discussions of some of the issues involved see E. D. Domar, 'Depreciation, Replacement and Growth, and Fluctuations," Economic journal, December 1957, pp. 655-658; E. D. Domar, 'Accelerated Depreciation: A Rejoinder," Quarterly Journal of Economics, May 1955, pp. 299-304; and R. Eisner, Accelerated Depreciation: Some Further Thoughts." Quarterly journal of Economics, May 1955, pp. 285-296. To illustrate deferred maintenance, the total number of train derailments reported to the FRA Bureau of Railway Safety rose 110 percent between 1961 and 1970, whereas the number of those derailments attributed to defects or failures or roadway alone rose to 315 percent during the same nine-year period. II gross capital expenditures are said to equal actual depreciation, then actual net investment in transportation property is, by definition, unchanged. The net investment in 1950, $24,592 million, is used in calculating the adjusted rate of return in the last column of Table 12. These fixed charges are in addition to and subordinated to payments for hired equipment, which have more than doubled from $376 million to $816 million over the same period. S John R. Meyer and Alexander 1. Morton 500 Payment of dividends in excess of net income draws down sharehiders' equity, Shareholders' equity declined from a high of $18.2 billion in 1966 to $16.6 billion in 1971. In which case, of course, the olc assets should be depredated or written off more rapidly in the accounts. The tone of the Conglomerate Merger Studies conducted by the ICC in 1969 that the ICC will scrutinize closely any diveion of funds generated in railroad operations to non-rail activities, The Conglomerate Merger Studies of the ICC are reprinted in Failing Railroads, Hearings Before the Senate Commerce Cornmitter, November 1970, Part 3, pp. 795 et seq. As an industry, railroading is characterized by strong feelings of pride and tradition an unusual degree of attachment to the physical property, and by a rather unique historkal involvement with the regions they serve, all of which contribute to a reluctance to disinvest or diversify out of railroading. This reluctance is likely to be amplified railroad managements that have ascended to their positions through the Onerating divisions of the railroad and have therefore been clo to the traditions of the rilroJ The amounts in this table do not include external capital in the form of fright cars supplied by private-car lines. The net investment in railroad cars by private-car owners has risen from $300 million in 1947 to $3.2 billion in 1971. (See ICC, 1971 and ICC, 1947 in Note 40.) "The equipment trust generally reqi es a 20 percent downpaynierit and a l5-year payout. It is sold on con1p itive bid and approval of issuance is required of the ICC, The condftionaJ sale contract is used not only for new equipment but for financing the rebuilding of equipment, where no downpayments are involved, or where more than the customary 15 years is desired. Conditional sale contracts are placed privately and involve no placement cost. ICC approval is not required." (Paul H. Banner, "Capit?l and QJtpot in the Railroad Industry," Papers of the Transportation Research Forum, 1968, pp. 137-138.) Banner, 1968 (See Note 54), pp. 137-138. It is believed that only '"ie railroad equipment mortgage has been defaulted on. In that case, involving the Florida East Coast line, the equipment was readily placed elsewhere, confirming the liquidity 01 the collateral. Of course, if defaults on equipment trusts, conditional sale contracts, and leases should become more widespread in the rail industry, the liquidity 01 this collateral could be considerably lessened. On the other hand, "the obligations of the industry for other forms of debt and equity rest upon the earning por of the firm since investment funded very little can be reposseserJ and courts have not looked with favor upon the disnembermesit of a railroad and public sale of its parts to satisfy debt. Rather, the company in ban&ruy continues in operation with bondholders receiving no return. Ultimately reorganization and recapitalizatmon occurs with possible loss to the bondholders." (Banner, 1968 (see Note 54), p. 138.) The real cost of funds provided through equipment debt or leasing is not so cheap te bond holders and equity holders as the nominal forms of financing is employcri, the claims of bondrate implies. Each time one of these holders and stock holders aie fu,ther subordinated. This may be a very real price that these particirlady if the marginal rate of return on the equipment junior creditors are paying, so financed is below the ctxt of this capital, as the decline in terest coverage ratios suggests to have been the case. There is an interesting debate whether the ready availability of external hands or equipment but the relatively unavailability of external funds for investment in road aid structures (for lack of collateral) biases the type of investrnen, macia in the rail indisiy. (See Banner, 1968 (see Note 54), pp. 137 if.) F OrIon quity. lion The U.S. Railroad Industry in the Post-World War II Period: A Profile 501 For purposes of Figure 1. operating ''expenses'' are defined as railway operating resenues minus profits and federal income taxes. Capital costs or rents do not intlucle aiiy i.ut fiji equity capital. more For suggestions and policy recommendations cn how this might be achieved, see Improving Railroad Productivity, Final Report of the Task Force on Railroad Productivity, Rgests road C are ilitlee, On, an lolical flee to ed in rating road t cars WflCrs ICC, 5-year ihe g the an the y and aland ' PP. ilroad a East of the arid I this unded an be of a uptcy zation 8 (see 'ap to these urther Jying, e coSt case. Is for d and ustry \Vashington, D.C.: National Commission on Productivity. November I 973